1. Diss. ETH No. 21627
Investigating new chiral 1,2-disubstituted ferrocenes
A dissertation submitted to
ETH ZURICH
for the degree of
DOCTOR OF SCIENCES
presented by
PETER ELADIO LUDWIG
Master of Science ETH in Chemistry
born on 5 January, 1984
citizen of Ardez GR and Spain
accepted on the recommendation of
Prof. Dr. Antonio Togni, examiner
Prof. Dr. Christophe Copéret, co-examiner
Zurich 2013

3. Über den Hügeln
lautlos der rote Milan
flieh seinen Schatten
– Hannes Joss
7th
August, 2013

5. Dedicated to
the memory of
my father

7. Acknowledgments
I would like to thank the people, who helped and supported me during my Ph.D. studies:
First of all I thank Professor Antonio Togni for letting me join his group and supervising
me during my thesis, for always having an open door, for the freedom he granted in realising
my own ideas, as well as for his support during my master thesis at Imperial College.
I thank Professor Christophe Copéret for the co-examination of this thesis and helpful
comments.
Special thanks go to Dr. Jan Welch for all the good advice while writing this thesis, and
for the proof–reading, also to Danny Rafaniello for designing the cover.
Also, I want to thank all of the students that I supervised during my thesis, including
Daniel Bachmann who did his master thesis with me, my semester students Johannes Boshkow,
Lucia Meier and Patrick Stücheli, and Luciano Mastrobuoni and Manuela Meister, who both
were my SiROP students.
Furthermore, for technical support during my research I would like to thank the fol-
lowing people: Oliver Sala for the DFT–calculations. Dr. Heinz Rüegger, Dr. Aitor Moreno,
Dr. René Verel and especially Barbara Czarniecki for NMR support. I also want to thank our
crystallography team, first of all my ’Hof–Kristallographen’ Dr. Rino Schwenk and Lukas Sigrist,
as well as the rest of the team: Dr. Raphael Aardoom, Dr. Katrin Niedermann, Dr. Michael Wörle
and Elli Otth.
I want to thank Professor Antonio Mezzetti, Dr. Pietro Butti, Dr. Jonas Bürgler, Dr. Michelle
Flückiger, Dr. Raffael Koller, Dr. Kyrill Stanek and Dr. Jan Welch for all their good advice at
the beginning of and throughout my thesis. In addition I thank all the current and former
members of the Togni and the Mezzetti group for all the fruitful discussions and the good
times together. I especially thank all my labmates from H230 over the years of whom I
would like to particularly mention Dr. Ján Cvengroš, Barbara Czarniecki, Rima Drissi, Takuya
Kamiyama, Raul Pereira, Dr. Raphaël Rochat, Dr. Amata Schira, Dr. Rino Schwenk, Lukas Sigrist
and of course once more my long–time table neighbour Dr. Jan Welch. For all the support,
hanging–out, cheering up and great activities outside of the lab I want to thank Dr. Raphael
Aardoom, Barbara Czarniecki, Rima Drissi, Dr. Michelle Flückiger, Alex Lauber, Dr. Esteban
Mejía, Dr. Katrin Niedermann, Dr. Tina Osswald, Dr. Raphaël Rochat, Dr. Nico Santschi, Dr. Rino
i

8. Schwenk, Remo Senn and Lukas Sigrist. Moreover, I would like to thank all the staff at ETH
Zurich that are doing a great job, most of all Guido Krucker.
I would also like to thank all the people that played an important role in my education
and were not just teachers or supervisors to me, but also Mentors and eventually became
friends: Hannes Joss, Dr. Rita Oberholzer, Karl Ehrensperger, Maurice Cosandey, Jochen Müller,
Dr. Daniel Stein, Dr. Alexander Ossenbach and Professor Susan E. Gibson.
A very special thank you goes to my father Peter Gaudenz Ludwig who imparted to me
his curiosity about the world and established the basis for my scientific career. I would also
like to thank my dear friend Thomas Rast who lived this curiosity with me especially during
our childhood years and my godfather Eduard Hunziker who fuelled my eagerness to learn by
introducing me to the world of computers and electronics.
Last and mostly, I want to thank my whole family and all of my friends for their sup-
port, especially during the rough times, and I want to give a special thank you to my mother
Agustina and my sister Alexandra, os quiero mucho.
ii

9. iii

10. iv

11. Contents
Abstract x
Zusammenfassung xii
1 Introduction 1
1.1 Ferrocenes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1.1 Nomenclature of enantiomerically pure 1,2-substituted fer-
rocenes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1.2 Synthetic routes towards enantiomerically pure 1,2-
substituted ferrocenes . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Aim and course of this Thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2 Synthetic approaches towards PSiP-Pigiphos 17
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.1.1 Attempts to improve the Ni(II)-Pigiphos system . . . . . . . . . . 19
2.1.2 Silyl ligands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.1.3 Aim of the project . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.2 The three fundamental approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.2.1 Nucleophilic silicon . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.2.2 The hydrosilylation route . . . . . . . . . . . . . . . . . . . . . . . . 23
2.2.3 The Umpolung . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.3 Synthetic Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.3.1 Hydrosilylation attempts . . . . . . . . . . . . . . . . . . . . . . . . 24
v

12. 2.3.2 Attempts towards an Umpolung . . . . . . . . . . . . . . . . . . . . 25
2.3.3 Umpolung via the thioacetal . . . . . . . . . . . . . . . . . . . . . . 25
2.3.4 Synthesis and characterisation of PSi derivatives of dithiane 26 30
2.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3 Synthetic approaches towards a chiral PSiP-Pincer ligand 35
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.1.1 Pincer ligands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.1.2 Pincer-like PSiP-ligands . . . . . . . . . . . . . . . . . . . . . . . . . 36
3.1.3 Aim of the project . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
3.2 Synthetic strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
3.3 Synthetic challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
3.3.1 Synthetic approach towards the PSiP-pincer like ligand 3 . . . . 38
3.3.2 Synthetic approach towards the PPP-pincer analogue 4 . . . . . 39
3.3.3 Explanation for the synthetic difficulties . . . . . . . . . . . . . . . 40
3.4 The sulfoxophosphine ligand 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
3.4.1 Structure discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
3.4.2 Complexation Experiments . . . . . . . . . . . . . . . . . . . . . . . 43
3.4.3 Catalytic experiments . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
4 Biferrocenylsulfoxides and Biferrocenylsulfides 51
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
4.1.1 Sulfoxide ligands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
4.1.2 Known Biferrocenyl compounds . . . . . . . . . . . . . . . . . . . . 56
4.2 Synthesis and structural features of BiFeSO 6 . . . . . . . . . . . . . . . . . . . . . . 60
4.3 Computational experiments with regard to the rotational barrier of BiFeSO 6 . . 63
4.3.1 Computational details . . . . . . . . . . . . . . . . . . . . . . . . . . 63
4.3.2 Computational results . . . . . . . . . . . . . . . . . . . . . . . . . . 63
vi

13. 4.4 Synthesis and structural features of BiFeS 7 . . . . . . . . . . . . . . . . . . . . . . . 71
4.5 X-ray structure of BiFeSO 6b and BIFES 7 . . . . . . . . . . . . . . . . . . . . . . . . 72
4.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
5 Side projects 75
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
5.2 Acidity of [Ni(II)-(Pigiphos)L]2+
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
5.2.1 Fluoride Ion Affinity . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
5.2.2 Aim of the project . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
5.2.3 Synthesis of [fluoro-Ni(II)-(Pigiphos)]+
. . . . . . . . . . . . . . . 76
5.2.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
5.3 Towards a chiral ferrocenyl building block . . . . . . . . . . . . . . . . . . . . . . . . 80
5.3.1 The bromo stannyl ferrocene . . . . . . . . . . . . . . . . . . . . . 80
5.3.2 The Bromo phosphino ferrocene . . . . . . . . . . . . . . . . . . . 82
5.3.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
5.4 Ferrocenyl-(trifluoromethyl) sulfide . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
5.4.1 Synthetic approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
5.4.2 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
6 Conclusion and Outlook 87
6.1 PSiP-Pigiphos . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
6.2 PSiP-pincer like ligand . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
6.3 Biferrocenylsulfoxide and Biferrocenylsulfide . . . . . . . . . . . . . . . . . . . . . . 88
6.4 Side Projects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
6.4.1 Acidity of [Ni(II)-PigiphosL]2+
. . . . . . . . . . . . . . . . . . . . . 88
6.4.2 Ferrocenyl building block . . . . . . . . . . . . . . . . . . . . . . . . 88
6.4.3 Ferrocenyl-(trifluoromethyl) sulfide . . . . . . . . . . . . . . . . . 89
6.5 General outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
7 Experimental 91
vii

14. 7.1 General Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
7.1.1 Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
7.1.2 Chemicals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
7.1.3 Analytical Techniques and Instruments . . . . . . . . . . . . . . . 92
7.2 Syntheses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
7.2.1 Ligands and Complexes . . . . . . . . . . . . . . . . . . . . . . . . . 93
7.2.2 Substrates and Catalyses . . . . . . . . . . . . . . . . . . . . . . . . 115
References 119
8 Appendix xiii
8.1 Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii
8.2 Crystallographic Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv
8.3 Curriculum Vitae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xix
viii

15. ix

16. Abstract
This dissertation reports investigations of new chiral 1,2-disubstituted ferrocenyl compounds,
with respect to their synthesis, properties and applications.
A synthetic approach to a PSiP-Pigiphos analogue 1 was explored. Due to steric hindrance
and synthetic challenges encountered, a double substitution of silicon, by a bulky ferrocenyl
moitety was unsuccessful. Nonetheless, the approach lead to the synthesis of a PSi ligand (2),
which underwent Si–H activation with platinum(0) to yield a hydridoplatinum(II) complex
(cf. Scheme 1).
Fe
Si
S S
P Pt
Ph2 PPh3
H
Fe
Si
SS
PPh2 H
[Pt(PPh3)4]
2
Scheme 1: Si–H activation of PSi ligand 2 with [Pd(PPh3)4].
Due to the problems encountered in the PSiP-Pigiphos synthesis, the synthesis of an alternative
PSiP-pincer like ligand 3, that would form five membered metallacycles upon Si–H activation
was investigated. At the same time, the synthesis of a PPP analogue 4 was attempted, leading
to bissulfoxophosphine 5 as an intermediate compound (cf. Scheme 2).
Fe
Fe
R
E
S S
toltol
O
OECl2RFe
PPh2
S
Ph
O
LDA
3: E = SiH
4: E = P
Fe
Fe
R
E
P P
Ph2Ph2
t-BuLi
ClPPh2
5: E = P, R = Ph
Scheme 2: Attempted synthesis of PSiP ligand 3 and its PPP analogue 4.
The bissulfoxophosphine 5 formed complexes with palladium(II), platinum(II) and
rhodium(I), with the partially characterised rhodium complex showing some activity in the
x

17. Miyaura-Hayashi reaction with low enantiomeric excess (16 %ee) and the structurally charac-
terised palladium complex showing activity in allylic substitution with enantiomeric excess up
to 78 %ee. The synthesis of PSiP ligand 3 and PPP ligand 4 both failed during the coupling
of the ferrocenyl moieties to the central donor atom, most probably due to a oxygen transfer
from the sulfoxide moiety to the eletrophile used in the synthesis.
The synthesis of a bis(ferrocenylsulfoxide) (BiFeSO) 6 was also developed (cf. Scheme 3),
resulting in two compounds that are apparent atropisomers of each other. One of the prod-
ucts 6b was fully characterised, including an X-ray structure. On basis of this data, quantum
chemical calculations were performed to test the atropisomery hypothesis. First the energy
necessary for a configurational change from 6b to 6a was calculated. The values obtained
from that calculation suggest that a configurational change would not take place at rt, there-
fore supporting the concept of two atropisomers. Furthermore, the 1
H-NMR spectra for the
suggested configuration of 6a and the known configuration of 6b were calculated. The results
were in good agreement to the observed 1
H-NMR signals of the ferrocenyl protons of the two
compounds. 6b was reduced to give the bis(ferrocenylsulfide) (BiFeS) 7 (cf. Scheme 3). While
complexation experiments with BiFeSO 6 was unsuccessful, BiFeS 7 seemed to eliminate iso-
butene, when reacted with mercury(II)bromide.
Fe
S t-Bu
O
Li
1. CuCN
2. O2
Fe
Fe
S
S
t-Bu
O
t-Bu
O
NEt3
SiHCl3
Fe
Fe
St-Bu
t-BuS
6 7
Scheme 3: Synthesis of BiFeSO 6 and BiFeS 7.
To complement the primary aims, three side projects were also undertaken during the course
of this work. With the estimation of the fluoride ion affinity of [Ni(II)-(Pigiphos)L]2+
in mind,
the synthesis of [fluoro-Ni(II)-Pigiphos]+
tetrafluoroborate 8 was developed. Due to the prob-
lems encountered in the synthesis of the PSiP ligand 3 and its PPP analogue 4 the synthesis
of bromo-2-(tri-n-butylstannyl)ferrocene 9 was developed with the intention to obtain an ‘in-
ert’ chiral building block, to be able to circumvent problems caused by the sulfoxide moiety. A
third side project lead to the synthesis of (Trifluoromethyl)ferrocenylsulfide 10 using the Togni
acid reagent, with the initial intention to synthesise a less rigid BiFeSO type compound.
xi

18. Zusammenfassung
Die vorliegende Dissertation befasst sich mit der Erforschung neuer chiraler 1,2-
disubstituierter Ferrocenverbindungen mit Bezug auf ihre Darstellung, Eigenschaften und An-
wendungsmöglichkeiten.
Die Darstellung eines PSiP Analogs 1 zu Pigiphos wurde untersucht. Die doppelte Substitution
am Silizium war nicht erfolgreich, da die einzige stabile Ausgangssubstanz für diesen Schritt
eine zu hohe sterische Hinderung aufwies. Stattdessen wurde ein PSi ligand 2 dargestellt,
welcher durch Si–H-Aktivierung mit Platin(0) einen Hydridoplatin(II) Komplex bildet (vgl.
Schema 1).
Fe
Si
S S
P Pt
Ph2 PPh3
H
Fe
Si
SS
PPh2 H
[Pt(PPh3)4]
2
Schema 1: Si–H des PSi-Liganden 2 mit [Pd(PPh3)4].
Da die Darstellung von PSiP-Pigiphos ohne Erfolg blieb, wurde die Synthese eines alternativen
pincerartigen PSiP-Liganden 3 untersucht, welcher durch Si–H-Aktivierung zwei füngliedrige
Metallacyclen bilden würde. Gleichzeitig wurde die Synthese eines PPP-Analogs 4 untersucht,
wobei Bissulfoxophosphin 5 als Zwischenprodukt gewonnen wurde (vgl. Schema 2).
Fe
Fe
R
E
S S
toltol
O
OECl2RFe
PPh2
S
Ph
O
LDA
3: E = SiH
4: E = P
Fe
Fe
R
E
P P
Ph2Ph2
t-BuLi
ClPPh2
5: E = P, R = Ph
Schema 2: Versuchte Darstellung des PSiP-Liganden 3 und seines PPP-Analogs 4.
Bissulfoxophosphin 5 bildet Komplexe mit Palladium(II), Platin(II) und Rhodium(I). Während
xii

19. der nicht charakterisierte Rhodiumkomplex in der Miyaura-Hayashi-Reaktion nur einen gerin-
gen Enantiomerenüberschuss (16 %ee) erzeugte, konnte mit dem Palladiumkomplex in einer
allylischen Substitution ein Enantiomerenüberschuss von bis zu 78 %ee erreicht werden. Die
Synthesen des PSiP- 3 und PPP-Liganden 4 scheiterten beide während der Kupplung der Ferro-
cenyleinheiten an das zentrale Donoratom, wahrscheinlich aufgrund eines Sauerstofftransfers
der Sulfoxidgruppe auf das in der Synthese eingesetzte Elektrophil.
In einem weiteren Schritt wurde die Synthese von Bis(ferrocenylsulfoxide) (BiFeSO) 6 real-
isiert (vgl. Schema 3). Diese lieferte zwei scheinbar atropisomere Produkte. Eines der Pro-
dukte 6b konnte inklusive einer Röntgenstrukturanalyse vollständig charakterisiert werden.
Mit den daraus gewonnenen Daten wurden quantenchemische Berechnungen durchgeführt,
um die Atropisomeriehypothese zu stützen. Zunächst wurde die nötige Energie für einen Kon-
figurationswechsel von 6b zu 6a berechnet. Die daraus berechneten Energien legen nahe,
dass ein Konfigurationswechsel bei Raumtemperatur nicht stattfindet und entsprechend von
Atropisomeren ausgegangen werden kann. Weiter wurden 1
H-NMR Spektren von der für 6a
berechneten Struktur, wie auch der bekannten Konfiguration von 6b berechnet, wobei die
Ergebnisse in Übereinstimmung mit den gemessenen 1
H-NMR Spektren sind. 6b wurde in
einem weiteren Schritt zu Bis(ferrocenylsulfid) (BiFeS) 7 reduziert (vgl. Schema 3). Während
Komplexierungsversuche mit BiFeSO 6 keinen Erfolg brachten, schien BiFeS 7 bei der Umset-
zung mit Quecksilber(II)bromid iso-Buten zu eliminieren.
Fe
S t-Bu
O
Li
1. CuCN
2. O2
Fe
Fe
S
S
t-Bu
O
t-Bu
O
NEt3
SiHCl3
Fe
Fe
St-Bu
t-BuS
6 7
Schema 3: Darstellung von BiFeSO 6 und BiFeS 7.
Zur Ergänzung der Hauptprojekte, wurden drei Nebenprojekte verfolgt. Aus der Ab-
sicht die Fluoridionenaffinität von [Ni(II)-(Pigiphos)L]2+
zu bestimmen, wurde die Syn-
these von [Fluoro-Ni(II)-Pigiphos]+
tetrafluoroborat 8 realisiert. Aufgrund der Probleme bei
der Darstellung des PSiP- 3 und PPP-Liganden 4 wurde die Synthese vom Bromo-2-(tri-n-
butylstannyl)ferrocen 9 entwickelt. Dies in der Absicht einen “inerten”, chiralen Baustein zu
erhalten, um die von der Sulfoxidgruppe verursachten Probleme zu umgehen. In einem drit-
ten Nebenprojekt wurde die Synthese von (Trifluoromethyl)ferrocenylsulfid 10 mit dem Togni-
Säure-Reagenz entwickelt, in der Absicht ein weniger starres BiFeSO-Derivat darzustellen.
xiii

20. xiv

21. 1 Introduction
1.1 Ferrocenes
Since its nearly simultaneous discovery by Keally and Pauson[1]
and Miller et al.[2]
in 1951,
ferrocene has been found to be a versatile component of chemical compounds that find ap-
plications in many different chemical areas ranging from homogeneous catalysis to material
sciences and biochemistry.[3–5]
Its stability, which arises from its aromaticity,[6]
paired with
its three dimensional structure[7–9]
makes it an ideal backbone for ligands used in asymmetric
catalysis.[5,10,11]
1.1.1 Nomenclature of enantiomerically pure 1,2-substituted ferrocenes
Unlike the planar benzene homoannular disubstituted ferrocenes bearing to different sub-
stituents do not have a mirror symmetry and are therefore chiral.[12,13]
The absolute configu-
ration of such planar chiral ferrocenes, is assigned following the rules proposed by Schlögel in
1967.[14]
Assignement of the absolute stereochemical configuration of 1,2-disubstituted fer-
rocenes is made by looking along the C5 axis of the ferrocene from the side of the more highly
substituted Cp-ring and arranging the substituents on that ring by their Cahn, Ingold, Prelog
priorities.[15–17]
The absolute configuration (R) or (S) can thus be assigned depending on the
clockwise or counterclockwise, respectively, nature of the resulting sequence of substituents
(cf. Scheme 4). If there are more than three substituents attached to the ring, only the three
with the highest priority are taken into consideration.
Fe
R2
R1
C5 axis
S configuration assuming
R1 has higher priority than R2
(a)
(b)
Fe
Fe
S
S
O
O
(RS,RS,Ra,RFc,RFc)-Bis-[2-(t-butylsulfinyl)ferrocene]
Scheme 4: Assignment of chiral planar configuration following Schlögel’s rule.
1

22. 1 Introduction
In order to distinguish planar chirality from other chirality units, such as central or axial chi-
rality, present in a molecule a "p" subscript is often used next to the assigned configuration.
The use of an "Fc" subscript has also become more common in ferrocene chemistry, in order
to avoid confusion with stereogenic phosphorus atoms in molecules, for which a "P" subscript
is often used. Conventionally, chiral elements have the following priorities: central > axial >
planar (cf. Scheme 4).[18]
1.1.2 Synthetic routes towards enantiomerically pure 1,2-substituted ferrocenes
Various methods have been developed to introduce planar chirality to ferrocenes, which, in
principle, can be divided into three types: A) diastereoselective directed ortho-metalation,
B) enantioselective directed ortho-metalation and C) resolution of racemic planar chiral
ferrocenes (cf. Scheme 5).[18]
In case A, a chiral auxiliary is used as a chiral directing metalation group (DMG). The
auxiliary has the ability to coordinate organolithium or lithium amide species and, through
the complex induced proximity effect (CIPE),[19]
is therefore able to diastereoselectively
deprotonate one ortho position on the ferrocene. The resulting lithium ferrocene can then be
quenched with an electrophile to yield a planar chiral 1,2-disubstituted ferrocene. In order
to introduce planar chirality through CIPE, the auxiliaries feature nitrogen or oxygen lone
pair coordinating sites. In contrast, in case B, the DMG is achiral and the method relies on
chiral lithiation agents to differentiate between the prochiral ortho positions. For method C,
on the other hand, the racemate is first synthesised and later kinetically resolved, either by
enzymatic or non-enzymatic kinetic resolution.
As type A is the most developed of the methods discussed and has also been the basis for the
work described in this thesis, deeper discussion of work done using this method will follow,
while methods B and C will be discussed briefly in this section.
Early work based on enantioselective directed ortho-metalation (method B) used (–)-sparteine
on isopropylferrocene resulting in slight enantiomeric excess of 3 % ee.[20]
Work done by
Price et al. using a chiral lithium alkyl amide on ferrocenyldiphenylphosphinoxide[21]
resulted
in only moderate enantiomeric excesses (54 % ee).[21]
The first satisfactory results based on
method B were reported almost concurrently by Tsukazaki et al. by using (–)-sparteine for the
n-BuLi mediated lithiation of N,N-diisopropyl ferrocenecarboxamide with an enantiomeric
excess of up to 99 % ee (cf. Scheme 6).[22]
In addition to further reports of (–)-sparteine
mediated ortho-lithiation,[23,24]
more recent work by Dixon et al. also shows the effective use
of sparteine surrogates for enantioselective ortho-lithation.[25]
2

23. 1.1 Ferrocenes
Fe
DMG
A diastereoselective directed ortho-metallation
RLi
Fe
DMG
Li E+
Fe
DMG
Fe
DMG
E
E
or
enantiomerically pure
Fe
DMG
B enantioselective directed ortho-metallation
RLi / chiral diamine
Fe
DMG
Li E+
Fe
DMG
Fe
DMG
E
E
or
achiral
chiral lithium amide
Fe
R2
C kinetic resolution
kinetic resolution
racemate
R1
Fe
R1
R2
+
Fe
R2
A
Fe
R1
R2
+
Fe
R2
R1
Fe
R1
A
+
or
Scheme 5: Three principle methods to introduce planar chirality to ferrocenes.[18]
Fe O
N
i-Pr
i-Pr
1. 1.2 equiv n-BuLi / (–)-sparteine
Et2O, –78°C
2. Ph2CO
Fe O
N
i-Pr
i-Pr
CPh2
OH
91% yield
99%ee
Scheme 6: An example for method B as reported by Tsukazaki et al.[22]
3

24. 1 Introduction
The first use of kinetic resolution on planar chiral 1,2-disubstituted ferrocenes was Horeau’s
method[26–28]
as applied by Falk and Schlögl in order to determine the absolute config-
uration of (+)-1,2-(α-ketotetramethylene)-ferrocene,[29]
which they isolated by reaction
with (–)-menthylhydrazide followed by multiple recrystallisations.[30]
However, Horeau’s
method represents an analytical tool, rather than a useful synthetic method, as racemic
phenyl butyric acid is reacted with an enantiopure substance in order to determine the
enantiomeric excess of the unreacted phenyl butyric acid. Although stochiometric kinetic
resolution of planar chiral ferrocenes is still a topic of current investigation,[31]
a more elegant
method of kinetic resolution of planar chiral ferrocenes for synthetic purposes is of a catalytic
fashion. One way to achieve this is through enzyme-catalysed asymmetric reactions that
have a long history in a variety of applications.[32]
First investigations of this method were
reported in the late 1980s using baker’s yeast, while later work focused on the esterification
of 1,2-disubstituted ferrocenyl alcohols by lipase (cf. Scheme 7), giving up to 95 % ee at
32 % yield in case for Candida cylindracea lipase[33]
(for a list of examples of enzymatic kinetic
resolution see Deng et al.[18]
and references therein).
Fe
OH
CCL, vinyl acetate
N
Fe
N
OH
+ Fe
OH
N
Fe
N
OAc
+
32% yield
95%ee
42% yield
92%eeCCL = Candida Cylindracea Lipase
Scheme 7: Example for enzymatic kinetic resolution as reported by Lambusta et al.[33]
A potential alternative to the enzymatic resolution is represented by the use of asymmetric
catalysis for kinetic resolution. This method was first applied in 2006 by Bueno et al. using
Sharpless asymmetric dihydroxylation.[34]
In the same year, Ogasawara et al. reported a
kinetic resolution based on asymmetric ring closing metathesis (cf. Scheme 8), which became
a matter of further investigation in his group.[35–37]
4

25. 1.1 Ferrocenes
Fe
(R)-Mo cat.
0.005 mol/l in benzene
50°C, 24h
t-But-Bu
Fe
t-But-Bu
Fe
t-But-Bu
+
+
Fe
t-Bu
t-Bu
2
47% yield
95%ee
46% yield
96%ee
3% yield(R)-Mo cat.
t-Bu
t-Bu
O
OMo
N
i-Pr
i-Pr
Me
Ph
Me
(rac)
Scheme 8: Asymmetric ring closing metathesis as reported by Ogasawara et al.[38]
1.1.2.1 Ugi-approach Although the first synthesis and isolation of (rac)-[1-
(dimethylamino)ethyl]ferrocene 11 was already reported in 1957 by Hauser and Lindsay,[39]
no special interest was given to this material until resolution with (R)-(+)-tartaric acid as
well as its use in diastereoselective ortho lithiation was reported by Ugi and co-workers.[40,41]
Due to the tertiary amine 11’s importance to the synthesis of chiral ferrocene derivatives, it
has become known under the trivial name Ugi’s amine.
Synthesis of Ugi’s amine. Many attempts towards the improvement of the synthesis of
optically pure Ugi’s amine have been reported,[42–48]
among which enzymatic methods[43,48]
as well as Corey-Bakshi-Shibata reduction[44–46]
proved to be applicable on a multi-kilogram
scale.[46,48]
However, the most widely used synthetic route is based on the synthetic route,
improved to limit the formation of vinyl ferrocene in the alcohol activation step, reported
by Ugi’s and co-workers in 1972[42,49]
(cf. Scheme 9). Resolution is still performed using
(R)-(+)-tartaric acid to crystalise the (S)-11 tartrate from methanol. The (R)-11 tartrate is
then recovered through evaporation of the mother liquor and recrystallisation from aqueous
acetone.[40,49]
5

26. 1 Introduction
Fe Fe
CH3COCl
AlCl3, DCM
O
LiAlH4
benzene
Fe
OH
HOAc
benzene
Fe
OAc
HNMe2
MeOH
Fe
NMe2
resolution
Fe
NMe2
Fe
NMe2
(S)-11 (R)-11(rac)-11
+
Scheme 9: Synthesis of Ugi’s amine.[40,49]
Use in synthesis of 1,2-disubstituted ferrocenes. Ugi and co-workers showed, that
treatment of (R)-11 with n-BuLi leads to a directed ortho-lithiation.[40,41]
This is due to the
interaction with the nitrogen lone-pair, which stabilises the lithium ion at one of the ortho
positions more favourably than the other. Inspection of the two possible diastereomers of
lithiated (R)-Ugi’s amine 12 reveals that (R,SFc)-12 is disfavoured due to the steric interaction
of the methyl group with the Cp -ring, whereas the (R,RFc)-12 diastereomer can be formed
without any steric hindrance. This interaction results in a diastereomeric ratio up to 96:4 dr
for the final products, as demonstrated by quenching with a variety of electrophiles[40]
(cf.
Scheme 10).
Fe
NMe2
(R)-12
Li
Fe
(S)-12
Li
NMe2 E+
Fe
NMe2
E
up to 96:4drsteric repulsion
Scheme 10: Selective ortho lithiation of Ugi’s amine.[40,41]
As the resulting products are diastereomers, separation of the major and minor product can
usually by achieved by flash column chromatography or crystallisation yielding the major di-
6

27. 1.1 Ferrocenes
astereomer in high purity. In a further step, the dimethylamino group of the ortho-substituted
Ugi’s amine 13 can be substituted by convertion to a leaving group, e.g. under acidic
conditions or by methylation of the amine. Ugi and co-workers reported that substitution of
the amine takes place with full retention.[50]
They stated that the reaction seems to follow
a non-classical SN 1-mechanism, in which the N–C bond is cleaved simultaneously with the
Fe–C bond to form a carbenium ion. As a matter of fact, the stabilising effect of ferrocene
on adjacent carbenium ions was already known and had been thoroughly investigated at the
time,[51,52]
leading to the conclusion that there is a significant interaction between iron and
the double bond formed during an elimination process, resulting in an 18 e−
configuration
of the formal Fe(III) centre.[53,54]
The masked carbenium ion 14 is then attacked in an exo
fashion by a nucleophile, resulting in retention of the configuration (cf. Scheme 11).
Fe
LG: e.g. HNMe2
+,NMe3
+,OAc
LG
H
Me
Fe+
H
Me
14
Nu-
Fe
Nu
H
Me
Scheme 11: Non-classical SN 1-mechanism for the substitution at the "benzylic" carbon.[53,54]
Due to these properties, Ugi’s amine is used a the starting material for a wide variety of
ferrocene-based ligands with central and planar chirality having applications in asymmetric
catalysis.[55–68]
Some of these ligands can be synthesised in a simple two step reaction from
Ugi’s amine, as in the case of Josiphos (cf. Scheme 12).
Fe
NMe2
Fe
NMe2
Fe
PCy2
PPh2 PPh2
(R)-11 (R,SFc)-PPFA (R,SFc)-Josiphos
1. n-BuLi,
THF, –78°C
2. ClPPh2
HPCy2
AcOH, 80°C
Scheme 12: Synthesis of Josiphos.
7

28. 1 Introduction
1.1.2.2 Sulfoxide approach A more recent approach towards the synthesis of chiral 1,2-
disubstituted ferrocenes is based on chiral ferrocenyl sulfoxides. Their use in diastereoselec-
tive ortho-lithiation was first reported in 1993 by Kagan and co-workers.[69]
The chiral ferro-
cenyl sulfoxides used for the directed ortho-lithiation are readily accessible through enantio-
selective oxidation of the sulfide[69–71]
or by nucleophilic attack of lithioferrocene on optically
pure sulfinates[72–75]
or thiosulfinates[76–79]
(cf. Scheme 13).
Fe
LiO
S
p-tol
O
S
S
O
Fe Fe
S Sp-tol
O O
Fe
S 1 equiv cumene hydroperoxide
1 equiv Ti(Oi-Pr)4
2 equiv (S,S)-diethyl tartrate
1 equiv H2O
Fe
S
O
Scheme 13: Synthetic routes to chiral sulfoxides.[69–79]
Use in synthesis of 1,2-disubstituted ferrocenes. Ortho-lithiation of ferrocenyl sulfoxides
is usually effected by addition of n-BuLi or LDA, depending on the other sulfoxide substituent
(cf. Scheme 14). Like the nitrogen lone-pair in the case of Ugi’s amine (vide supra), the oxygen
lone-pair of the sulfoxide facilitates ortho-lithiation, favouring the lithiated diastereomer with
the sulfoxide substituent anti to the ferrocene. Therefore, the two commonly used ferrocenyl
sulfoxides (RS)-t-butylferrocenylsulfoxide 15 and (SS)-p-tolylferrocenylsulfoxide 16 give
1,2-disubstituted ferrocenes with opposite planar chirality (cf. Scheme 14).
An advantage of p-tolylsulfoxide 16 over t-butyl sulfoxide 15 is the possibility to replace the
8

29. 1.1 Ferrocenes
Fe
S p-tol
O
LDA
Fe
S
p-tolOLi
TMSCl
Fe
S p-tol
OTMS
Fe
S t-Bu
O
Fe
S
t-Bu
O
MeI
Fe
S
t-Bu
O
Li Men-BuLi
Scheme 14: Diastereoselective ortho lithiation of sulfoxides.[75,77]
sulfoxide by another substituent through attack with either t-BuLi[75]
or PhLi[80]
forming
the corresponding sulfoxide and lithioferrocene species. Subsequent quenching of the
lithioferrocene with an electrophile gives access to a large variety of ligands[81]
(cf. Scheme
15).
Fe
S p-tol
OR
t-BuLi
t-Bu
S
p-tol
O
Fe
R
Li
E+
Fe
R
E
Scheme 15: Substitution of p-tolyl sulfoxide.[75]
1.1.2.3 Chiral acetal approach Another approach towards enantiopure 1,2-disubstituted
ferrocenes developed in Kagan’s group utilises the chiral acetal 17 and was reported by Riant et
al. in 1993.[82]
The methoxymethyl dioxane 17 is readily accessible from ferrocene by a three
step synthesis with an overall yield of 82 % (cf. Scheme 16). The (S)-(–)-1,2,4-butanetriol
needed for the synthesis of hydroxymethyl acetal 18 can be readily obtained by reduction of
(S)-(–)-malic acid with borane.[83]
Therefore, the approach is also economically viable. The
9

30. 1 Introduction
directing effect in ortho lithiation arises from the stabilising effect of the methoxy group and
one of the dioxane oxygen atoms, which chelate the lithium at the ortho position that leads to
Fe
O
(MeO)3CH
p-tolyl sulfonic acid
80°C
Fe
O
O (S)-(–)-1,2,4-butanetriol
camphor sulfonic acid
CHCl3, 4Å, rt
Fe
H
OO
OH
18, 85%
1. NaH, THF, 0°C
2. MeI Fe
H
OO
O
17, 97%
Scheme 16: synthesis of the chiral acetal 17.[84]
the most favourable chelation ring, resulting in the (S)-lithioferrocene 19 yielding the product
in a diastereomeric ratio of 99:1 dr[82]
(cf. Scheme 17). Most probably, the orientation of the
oxygen, which is not involved in the lithium chelation towards the iron moiety may have a
major impact on diastereoselectivity. In the case of (R)-19 this atom is positioned endo with
respect to the iron centre, whereas in (S)-19 it is oriented exo.[84]
It has also been shown, that
the directing effect is of kinetic origin, since the diastereomeric excess decreases significantly
if the reaction temperature is raised, with 95:5 dr at 0 ◦
C.[82]
The directing acetal can be removed by hydrolysis after planar chirality has been introduced.
The resulting enantiopure 2-substituted formylferrocene has proven useful for synthesis of a
large variety of chiral ferrocenyl compounds (cf. Scheme 18).
1.1.2.4 Oxazolines Enantiomerically pure ferrocenyl oxazolines are readily synthesised
from ferrocenylacyl chloride and the corresponding amino alcohol (cf. Scheme 19). The enan-
tiomerically pure amino alcohols can be generated through the reduction of amino acids,[85,86]
whereby a large variety of chiral oxazolines are accessible.
10

31. 1.1 Ferrocenes
H
Fe
H
OO
O
t-BuLi
Fe
O
O
OLi
Fe O
O
OLi
H
(S)-19
(R)-19
E+
Fe
H
OO
O
E
99:1dr
–78°C
Scheme 17: Diastereoselective ortho-lithiation of acetal 17.[82,84]
Fe Fe
P
Ph
Fe
Fe
N N
H
Fe
Fe
OH
HO
Scheme 18: Ligands synthesised by following the acetal approach.[87–90]
Directed ortho-lithiation of enantiopure ferrocenyl oxazolines has been performed by the
treatement of the oxazoline with n-BuLi or s-BuLi in ethers at –78 ◦
C giving a diastereomeric
excess up to 97:3 dr.[91–95]
An alternative experimental procedure using hexanes as solvent
and TMEDA gave an diastereomeric excess of >99:1 dr. This method was designed by Sam-
makia et al. in order to test their hypothesis for directed ortho lithiation.[93,94]
They proposed
that control of diastereoselectivity is derived from the steric interaction of the bulky group on
the oxazoline with the butyl group of the butyllithium, rather than the interaction with the
ferrocene. Therefore the stereo information would be imparted in the transition state of the
deprotonation of the ortho position (cf. Scheme 20). However, other factors that may influ-
ence the diastereoselectivity exist and they should still be taken into consideration.
11

32. 1 Introduction
Fe
Cl
O
OHH2N
R
1. , Et3N, CH2Cl2
2. a, b or c
a: TsCl, Et3N, cat. DMAP, CH2Cl2
b: SOCl2, 20% K2CO3 (aq.)
c: PPh3, CCl4, NEt3, CH3CN
Fe
O
N
R
Scheme 19: Synthesis of enantiomerically pure ferrocenyl oxazolines.[92,93,95]
Fe
O
N
R
BuLi
Fe
H
Li Bu
N
O
R
Fe
H
LiBu
N
O
R steric repulsion
major minor
Fe
O
N
R
Fe
O
N
R
E+
E
E
major minor
Scheme 20: Diastereoselective ortho lithiation of ferrocenyl oxazolines.[93,94]
Hydrolysis of the oxazoline could be considered as a feasible method to replace the oxazoline
by another functionality. However, the donor features of the oxazoline make it useful as coor-
dination site for complexation and thus render an exchange unnecessary for the synthesis of
chiral ligands. This is one of the major advantages of the oxazoline approach,[18]
as it gives ac-
cess to asymmetric bidentate ligands in only a single reaction step, complementing the already
large variety of oxazoline ligands[96,97]
with their ferrocene derivatives.
12

33. 1.1 Ferrocenes
1.1.2.5 Directing groups containing phosphorus A variety of aryl phosphine deriva-
tives have been shown to have an ortho directing effect upon metallation.[98–103]
The
diastereoselective ortho metallation utilising chiral ferrocenyl phosphine derivatives seems
somewhat obvious. However, only a few successful examples are known. One of these is the
ortho-magnesiation reported by Nettekoven et al.[104–106]
(cf. Scheme 21). A diastereoselective
excess of 97:3 dr in quantitative yield was achieved, using iodine as the electrophile.
Fe
P
O
R
Fe
Mg
O
P
Fe
Mg O
P
major minorsterical repulsion
(i-Pr)2NMgBr
I2
Fe
P
O
R
I
97:3dr
R:
Scheme 21: Diastereoselective ortho magnesation reprted by Nettekoven et al.
Another successful example is that of the oxazaphospholidine-oxide reported by Xiao and
co-workers,[107–109]
which undergoes diastereoselective ortho-lithiation with t-BuLi, giving
a diasteremeric excess of >99:1 dr in yields varying between 45 – 95 %, depending on the
electrophile. They also discovered, that the yield decreases significantly with the use of n-BuLi
as lithiating agent, due to reaction with the phosphorus moiety (cf. Scheme 22),[107]
which is
a general problem in directed ortho lithiation of phosphine derivatives.[103]
An example using
a P(III) instead of a P(V) phosphorus derivative was patented by Pfaltz et al.,[110]
who used a
borane protected phosphine bearing chiral amidites to yield 1,2-disubstituted ferrocenes with
99:1 dr (cf. Scheme 23).
13

34. 1 Introduction
Fe
P
O
O
N
Ph
1. t-BuLi, –78°C
2. E(X)
E(X) = Me(I), I(I), TMS(Cl), TES(Cl), Ph2CO, B(OMe)3, PR2(Cl)
Fe
P
O
O
N
PhE
>99:1dr
Fe
P
O
O
N
Ph
1. n-BuLi, –78°C
2. MeI
Fe
P
O
O
N
PhMe
>99:1dr, 33%
Fe
P
O
n-Bu
N
MeO
Ph
50%
Scheme 22: Diastereoselective ortho lithiation of oxazaphospholidine-oxide as reported by Xiao and
co-workers.[107–109]
Fe
P
BH3
N
N
OMe
OMe
1. s-BuLi, Et2O, –78°C
2. E(X)
Fe
P
BH3
N
N
OMe
OMe
E
E(X) = TMS(Cl), PPh2(Cl), Br(CF2CF2Br)
Scheme 23: Diastereoselective ortho lithiation as reported by Pfaltz et al.[110]
14

35. 1.2 Aim and course of this Thesis
1.2 Aim and course of this Thesis
The initial motivation behind this thesis was to improve the Ni(II)-Pigiphos system that
was developed in the Togni group. The main problems encountered with the dicationic
Ni(II)-Pigiphos system arose from its strong bonding not only to the substrate, but also to
coordinating solvents as well as, in the case of the Nazarov cyclisation, the product (for detail
cf. Section 2.1). As a consequence the catalyst gets poisoned during the reaction. In order
to facilitate the release of the product from the Ni(II) catalyst in the Nazarov cyclisation and
therefore facilitate the completion of the catalytic cycle, a new ligand design was propound
that lowers the lewis acidity of the catalytic system and therefore weakens the bond of the
metal at the active site. A silyl donor as central coordination site in the ligand would meet
this goal. First, the decreased charge of the complex would already have an impact on Lewis
acidity. In addition the silyl donor is a stronger σ-donor than the phosphine, which results
in further elevation of the energy levels of the orbitals involved in σ-bonding. In case of
a square-planar complex this concerns orbitals with a1g, b1g and eu symmetry, therefore
including dz2 (a1g) and dx2−y2 (b1g), which represent HOMO and LUMO of a square planar
complex (cf. Scheme 24). As a consequence the release of the weakest bound ligand should
HOMO
LUMO
increase of σ-donation
free metal
Scheme 24: Effect of σ-donation on the MO diagram of a square planar complex.
be facilitated resulting in a higher accessibility of the active site. Therefore, the synthesis of
a PSiP-Pigiphos 1 analogue and the comparison of the PPP- and PSiP-Ni(II)-Pigiphos systems
with respect to their properties and catalytic activity was the initial goal of this thesis (cf.
Chapter 2). As the synthesis of a PSiP-Pigiphos analogue was unsuccessful, a simplification of
the system to an alternative PSiP 3 and PPP 4 tridentate ligand was considered (cf. Chapter
3). Synthetic difficulties encountered in the coupling of the two ferrocene moieties to the
15

36. 1 Introduction
central donor atom made the isolation of the desired products unfeasible. Nonetheless,
a bis(sulfoxo)phosphine 5 was isolated as an intermediate in the attempted synthesis of
the PPP-pincer 4. This bis(sulfoxo)phosphine 5 formed κ2
-complexes with palladium(II),
platinum(II) and rhodium(I), which also showed asymmetric catalytic activity. This sparked
interest in sulfoxide ligands leading to the design and synthesis of the bis(ferrocenylsulfoxide)
6 (cf. Chapter 4).
Fe
Si
X
Fe
PPh2 Ph2P
R
Fe
Fe
Si
P P
H
R
Ph2Ph2
Initial PSiP-Pigiphos analogue
Fe
Fe
P
P P
R
Ph2Ph2
Bis(sulfoxo)phosphine
isolated as intermediate
Fe
Fe
Ph
P
S S
toltol
O
O
Fe
Fe
S
S
t-Bu
O
t-Bu
O
1
Simplified PSiP and PPP system
3 4
5
Focus on a pure
sulfoxide ligand
6
Scheme 25: General conceptual scheme.
16

37. 2 Synthetic approaches towards PSiP-Pigiphos
2.1 Introduction
Ferrocenyl-based ligands developed for application in asymmetric catalysis have a long history
within the Togni group. Besides the well-known bidentate phosphine ligand Josiphos, a variety
of different ferrocene-based ligands (cf. Scheme 26) have been created and studied by former
and current members of the Togni group. Among these is the tridentate phosphine ligand
Pigiphos, which was first synthesised by Pierluigi Barbaro[60]
following a straightforward two
step synthesis starting from commercially available Ugi’s amine (cf. Scheme 27).
Fe
N
N
R'
R''
PPh2
Fe
P
Fe
PPh2 Ph2P
Cy
Fe
Fe
N N
H
Fe
PCy2
PPh2
Fe
FeP
Cy
Josiphos
Pigiphos
Scheme 26: Selection of ferrocene based ligands synthesised in the Togni group.
Pigiphos readily forms complexes with a wide variety of late transition metals,[60,112–115]
whereby the first reported asymmetric catalysis with the ligand used a ruthenium(II)-Pigiphos
complex for transfer hydrogenation of acetophenone.[112]
Special interest has been taken in
the dicationic nickel(II)-Pigiphos complex, which was first synthesised and used for asymmet-
ric acetalisation by Barbaro.[113]
As a chiral lewis acid it was also used as a catalyst for hy-
droamination,[111,116]
hydrophosphination,[117,118]
Nazarov-cyclisation[119,120]
and 1,3-dipolar
17

38. 2 Synthetic approaches towards PSiP-Pigiphos
Fe
N
1. t-BuLi
2. ClPPh2
Et2O, -78 °C Fe
N
PPh2
CyPH2, TFA
AcOH, 80 °C
Fe
P
Fe
PPh2 Ph2P
Cy
Scheme 27: Two step synthesis of (R)-(SFc)-Pigiphos derivatives starting from Ugi’s amine.[111]
O
CO2R2
R3Ph
R1
20a-h 21a-h
i) [Ni(II)-Pigiphos](ClO4)2
in situ
THF, rt
ii) CH2Cl2, rt
O
R1
Ph
CO2R2
R3
Compound R1
R2
R3
Yield (%) ee (%)
21a Me Et TMPa
84 86
21b Ph Et TMPa
85 87
21c Me Et PMPb
32 71
21d Ph Et PMPb
96 83
21e Me Pr TMPa
80 82
21f Ph Pr TMPa
82 88
21g Me Bn TMPa
58 45
21h Me Npc
TMPa
no reaction n.a.
Reaction times for full conversion are 6 – 8 d for substrates having R3
= TMP
and 9 – 15 d for R3
= PMP. a
TMP = 2,4,6-trimethoxyphenyl. b
PMP = 4-
methoxyphenyl. c
Np = 1-naphtyl.
Table 1: Ni-catalysed Nazarov cyclisations of various dialkenyl ketones[120]
cycloaddition reactions. Despite this variety of applications of the dicationic Ni(II)-Pigiphos
complex the strong binding of the dicationic Ni(II)-Pigiphos complex to coordinating solvents
is a considerable problem, that leads to catalyst poisoning and therefore low TON. Similarly,
in case of the Nazarov-cyclisation the strong binding of Ni(II)-Pigiphos to the product in the
catalytic cycle , leads to low TON, as well as long reaction times due to low TOF (cf. Table 1).
18

39. 2.1 Introduction
2.1.1 Attempts to improve the Ni(II)-Pigiphos system
In order to overcome the above mentioned activity problems, the introduction of an N-
heterocyclic carbene (NHC) moiety as a replacement for the central phosphorus donor site
in the Ni(II)-Pigiphos system was undertaken in our group. NHCs display similar bonding
properties to trialkylphosphines,[121,122]
but with the benefit of being much stronger σ-donors
in most cases. Although the synthesis of the NHC bearing Pigiphos analogue 22 has been
performed successfully (cf. Scheme 28),[123]
it turned out to have major disadvantages due
to the flexibility of the system caused by the additional bridging carbon atoms between the
ferrocene and the carbene moiety. Not only were lower enantiomeric excesses observed,
but in most cases no advantages over the Pigiphos catalytic system could be discerned. In
addition to the above mentioned conformational flexibility, the NHC-Pigiphos derivative
also showed relatively weak coordination of the NHC moiety to metal centres. For example,
an extraordinarily long NHC-Pd bond of 2.040(12) Å[123]
is observed in the Pd(II) iodo
complex of this ligand. This unusually long distance between the donor ligand and metal is
most likely a result of disfavoured seven membered metallacycles formed by coordination
of the phosphine groups. As a consequence of these results and observations, an alternative
modification of Pigiphos was thought to be necessary.
Fe
NMe2 1. t-BuLi, Et2O
2. ClPPh2
3. AcOAc, 2-5 h,
100 °C
Fe
OAc
PPh2
1. Imidazole,
AcCN/H2O
2. NaI, EtOH,
3 h, rt Fe
Fe
N N
HPPh2 Ph2P
22
Scheme 28: Synthesis of the NHC-Pigiphos analogue 22.
19

40. 2 Synthetic approaches towards PSiP-Pigiphos
2.1.2 Silyl ligands
Although Wilkinson reported the first transition metal silyl derivative as early as 1956,[124]
the
developement of the field was initially slow.[125]
Only after the discovery of transition-metal-
catalysed hydrosilylation of alkenes[126]
and the importance of the Si–H activation by oxidative
addition behind it,[127]
did interest in the area start to grow. Silyl ligands are particularly
strong σ-donors and have been shown to have a strong trans influence. X-ray crystallographic
analyses show Pt–Cl bond lengths trans to the silyl donor are up to 0.161 Å longer than those
in PtCl2−
4
with Pt–Cl bond lengths of 2.465 (2) Å in case of the triphenylsilyl platinum com-
plex 23 (cf. Figure 1).[128]
This fact, together with the low frequency IR signals for ν(Pt–Cl)
at 239 cm−1[129]
observed are clear indicators of the strong trans influence of silicon donor
ligands.
PtCl
Si
P2
P1
Figure 1: X-ray structure of the triphenylsilyl platinum complex 23.[128]
Currently, there is a special interest in incorporating silyl donors into ancillary ligand frame-
works. In such a framework, the strong trans labilising σ-donor properties of the silyl donor
can be fully utilised.[130]
Such ligands form coordinatively unsaturated complexes and have
been purported to show enhanced reactivities.[131,132]
Many complexes of this type have been
reported and some have shown interesting catalytic activity.[132–144]
There are a variety of
methods to form Si–M bonds in a complex. Among the most common is Si–H activation. As
Si–H bonds are known to be more reactive toward oxidative addition than other Si–X bonds,
this represents one of the most viable paths to Si–M complexes. Methods using transition
metal anions or silyl anions have also been reported.[125]
20

41. 2.2 The three fundamental approaches
2.1.3 Aim of the project
Due to the problems encountered in catalysis with Pigiphos and the known properties of silyl
donors, it was assumed that reactivity, in terms of TOF, could be enhanced if a PSiP-Pigiphos
analogue could be synthesised and applied. The strong trans labilising effect of the silyl
donor should lead to an increased exchange rate at the active site as well as a weakening
of the product–catalyst complex. The only potential drawback of such an approach may be
the monocationic character of the Ni(II)-PSiP-Pigiphos complex formed, the Lewis acidity
of which might be lowered to the point at which it no longer activates the substrate. This
particular problem might be overcome by chosing a different metal-ligand system, thus
adding intrinsic value to the proposed PSiP ligand class. Therefore, the aim of this work is
to prepare a PSiP-Pigiphos analogue and complex it, by Si–H activation, to form a catalytic
system comparable to the Ni(II)-Pigiphos system discussed above.
Fe
Si
X
Fe
PPh2 Ph2P
R
X = H, Cl
R = Me, Ph
1
Scheme 29: Generalised structure of the proposed PSiP-Pigiphos ligand 1
2.2 The three fundamental approaches
To synthesise a PSiP-Pigiphos ligand 1 three different approaches were considered (cf. Scheme
30). Based on the known Pigiphos synthesis from Ugi’s amine (vide supra), a nucleophilic
silicon reagent would be most useful. Hydrosilylation of a vinyl ferrocene or Umpolung of the
"benzylic carbon" at the ferrocene would also be effective strategies for the formation of the
desired ligand systems.
21

42. 2 Synthetic approaches towards PSiP-Pigiphos
Fe
Si
X
Fe
PPh2 Ph2P
R
Fe
PPh2
Fe
PPh2
Fe
PPh2
LG
M
+ RSiH2X + RSiM2X
+ RSiCl2X
Hydrosilylation route Nucleophilic silicon route
Umpolung route
1
Scheme 30: Three fundamental retrosynthetic routes to synthesise PSiP-Pigiphos 1
2.2.1 Nucleophilic silicon
The simplest form of a nucleophilic silane moiety, is the analogue of the carbanion, which
here may be referred to as silicon anions for simplicity. As a matter of fact, silicon anions
have been the subject of investigation for the better part of the past century.[145–148]
Usu-
ally, symmetrically substituted disilanes are treated with alkali metals in ether solution to
give alkali silicides. Metallation of halosilanes has also been reported, whereby a disilane is
formed in a Würtz-coupling-type reaction which is then cleaved by the alkali metal. Finally,
deprotonation of certain silanes by potassium hydride has been observed as well.[149]
One
of the most common silyllithium compounds is triphenylsilyllithium, the reaction of which
with diphenylphosphinoacetylferrocene could provide a starting point for PSiP-Pigiphos, since
the phenyl substituents on silicon may be readily removed with triflic acid.[150]
The resulting
silyl triflate may be lithiated a second time leading to the desired product in a multistep syn-
22

43. 2.2 The three fundamental approaches
thesis (cf. Scheme 31). However, the harsh reaction conditions and multistep synthetic route
render such an approach a significant challenge.
Fe
Si
X
Fe
PPh2 Ph2P
R
Fe
PPh2
X LiSiPh3
Fe
PPh2
SiPh3
Fe
PPh2
SiPh2
HOTf
Lithiation
OTf
1
Scheme 31: Theoretical multistep synthetic route towards PSiP-Pigiphos using silyl lithium.
Rhodium(I) or copper(I) activated Si–B bonds may also act as silyl nucleophiles. Nucleophilic
silicon compounds of this nature form the corresponding silicon cuprate or rhodate in cat-
alytic quantities. To date, these metal-silicon compounds have been reacted with electrophiles
such as aldehydes or α,β-unsaturated carbonyls.[151,152]
This kind of reaction has only been
reported for monoborylsilanes. Therefore, this approach to the synthesis of PSiP-Pigiphos,
requiring boryl silanes, is also synthetically complicated, since the boryl silanes are synthe-
sised from corresponding chlorosilanes in a multistep process, hence resulting in a complex,
multistep synthesis of the desired product.
2.2.2 The hydrosilylation route
Since the first use of the term "catalytic hydrosilylation" by Ojima et al.[153]
many new cat-
alytic systems have been reported,[154,155]
and the method has been developed into one of the
most important uses of homogeneous platinum catalysis, second in importance only to the vul-
canisation of silicone rubber.[156]
Considering the ready accessibility of vinyl ferrocenes from
Ugi’s amine[157]
hydroslilylation may be a feasible synthetic strategy for a PSiP backbone. The
only foreseeable pitfall of this method may arise from anti-Markovnikov addition to the vinyl
group, which would lead to a C2 tether instead of a C1 tether between the silicon moiety and
the ferrocenyl unit.
23

44. 2 Synthetic approaches towards PSiP-Pigiphos
2.2.3 The Umpolung
As chlorosilanes are not only good electrophiles but are also commercially available in many
varieties, an Umpolung of the benzylic position of a ferrocene derivative might be a straight-
forward path towards the synthesis of a PSiP-Pigiphos. Different approaches towards such
an Umpolung may be considered. Although Gmelin reports the existence of ferrocenyl-
(chlorozirconocenyl)-methane,[158]
the original literature[?,159]
shows that, as one would ex-
pect, the hydrozirconation of formyl ferrocene using Schwartz’ reagent results in the zir-
conocene bound to oxygen, with the hydrogen adding to the adjacent carbon. However,
such an approach could be considered, as well as the potential hydrozirconation of a vinyl
ferrocene, despite the potential for the formation of a C2 tethered system.
A further approach would be a Corey-Seebach-Umpolung[160]
, which is a simple method for
the synthesis of acylsilanes.[161]
This method has already been demonstrated for formyl fer-
rocene by Reuter et al.[162]
By using an enantiomerically pure formyl ferrocenyl phosphine,
this seems a tantalisingly elegant approach. Kondo et al. described a synthetic route to fer-
rocenylmethyllithium in the early 1970s, through reductive lithiation of ferrocenylmethyl-
methoxide.[163,164]
Two decades later Knochel and co-workers claimed a similar approach to
stable α-ferrocenyllithium derivatives starting from α-thioethers, -ethers and -amines.[165]
In
this case, the approach via the amine is of interest since Ugi’s amine may be used to introduce
planar chirality (cf. Paragraph 1.1.2.1).
2.3 Synthetic Results
The different approaches towards the synthesis of PSiP-Pigiphos mentioned in the introduction
of this chapter were investigated in parallel to determine, as quickly as possible, which would
be the most feasible. Application of a nucleophilic source of silicon was ruled out during
preliminary investigations, due to the foreseen difficulties concerning harsh reaction condi-
tions combined with a long multistep reaction path (vide supra). The respective reaction paths
and the associated difficulties are discussed in more detail to clarify the choice of synthetic
approach.
2.3.1 Hydrosilylation attempts
Although hydrosilylation is a widely used method for introduction of silicon or a hydroxy group
into a molecule, its use on vinylferrocenes is largely unkown. The work by Morán et al. on oc-
takis(dimethylsiloxy)octasilsesquioxanes[166]
is often cited, as is Losada’s work on ferrocenyl
24

45. 2.3 Synthetic Results
functionalised silane based dendrimers.[167,168]
Both use Karstedt’s catalyst for the reaction
with tertiary silanes. Regrettably, their catalytic system failed to yield hydrosilylation products
when chlorophenylsilane or chloromethylsilane were used in combination with vinylferrocene
or (diphenylphosphino)vinylferrocene.
Other attempts using chloroplatinic acid, a known catalyst for the hydrosilylation of styrenes
by chlorosilanes,[169,170]
did not result in the desired product, regardless of substrate. Due to
the failure of these experiments to produce the target compounds and the fact that hydrosi-
lylation should lead to the less favoured C2-tethered product, efforts along this route were
ceased.
2.3.2 Attempts towards an Umpolung
An Umpolung strategy by reductive lithiation as described by Knochel and co-workers[165]
was
one of the first methods for the generation of silylated ferrocenyl materials pursued in this
work. Despite several attempts to follow this reaction protocol, the results proved unrepro-
ducible. A hydrozirconation approach, as described by Etiévant,[159,171]
was performed in two
test reactions with formyl ferrocene which indicated that starting materials were consumed,
but the products of these trials could not be isolated. Meanwhile, an Umpolung following
the Corey-Seebach route[162]
was successful and the focus of further experiments was directed
towards this particular approach.
2.3.3 Umpolung via the thioacetal
Initially, planar chirality was imparted to the ferrocene derivative, by following the reaction
protocol of Riant et al.[84]
The chiral acetal 18 was synthesised in a two-step reaction from
formyl ferrocene, by using (S)-(–)-1,2,4-butanetriol, which can be readily prepared from (S)-
malic acid by reduction with borane,[83]
to introduce stereochemical information. Deproto-
nation of the hydroxy group followed by methylation leads to the ether 17 which undergoes
selective ortho lithiation of the ferrocene moiety, when reacted with t-butyllithium. Quench-
ing the lithiated species of 17 with chlorodiphenylphosphine gave compound 24, which un-
derwent an acetal exchange in HCl-saturated benzene with 1,3-propanedithiol to give the
thioacetal 25 in an overall yield of about 13 % (cf. Scheme 32).
Recrystallisation of the thioacetal from DCM/n-hexane 25 gave single crystals suitable for X-
25

46. 2 Synthetic approaches towards PSiP-Pigiphos
Fe
O
H TSA
HC(OMe)3 Fe
H
O O
CHCl3, MS 4Å
HO OH
OH
Fe
H
O O
OH
NaH, MeI
THF
Fe
H
O O
OMe
1. t-BuLi
2. ClPPh2
Et2O Fe
H
O O
OMe
PPh2
1,3-propanedithiol, HCl
benzene Fe
H
S S
PPh2
99%
48%
82%82% 41%
25
18
17 24
Scheme 32: Multistep reaction path towards phosphinoferrocenyl-1,3-dithiane 25.
P1
C17
C11
C1 C5
S2
S1C23
Figure 2: X-ray structure of the phosphinodithiane 25.
ray crystallography (cf. Figure 2). In order to judge the importance of the structural features
from the X-ray structure of compound 25 the structural parameters were compared to those
for 1,1 -bis(1,3-dithian-2-yl)ferrocene 26 reported by Hartinger et al.[172]
as well as the 1,1 -
bis(diphenylphosphenyl)-2,2 -bis(1,3-dioxan-2-yl)ferrocene 27 reported by Connell et al.[173]
(cf. Scheme 33).
26

47. 2.3 Synthetic Results
Fe
H
S S
PPh2
Fe
H
S S
Fe
H
O O
PPh2
H
S S
H
Ph2P
O O
2526 27
Scheme 33: Compounds used for structural comparison. From left to right: dithioacetal 26,[172]
phosphinothioacetal 25 and bisdiphenylphosphino diacetal 27.[173]
The bond lengths of the substituents to the ferrocene are largely the same, with their bond
length differences within the experimental standard deviations calculated. In order to assess
the conformational differences between the structures, φ1 was defined to be the angle be-
tween the Cp-plane and the plane including the base of the trigonal pyramid formed by C(5),
S(1), S(2) and C(23), with C(23) being the apex of the pyramid. This angle was compared
with the angle between the planes passing through the respective atoms of compounds 26
and 27. Interestingly, in case of Hartinger’s bis(dithianyl)ferrocene, φ1 varies significantly be-
tween the two thioacetal groups, having the values of 76.4° and 84.2°. Regardless of the fact
that the values for the bis(dithianyl)ferrocene differ so much from each other, the value of φ1
for compound 25 is still significantly smaller at 66.4°, while the ferrocenyl acetal reported by
Connell shows angles for φ1 of 59.8° and 54.0° respectively (cf. Table 2).
Compound 26 [°] Compound 25 [°] Compound 27 [°]
76.4 66.4 59.8
84.2 54.0
Table 2: Angles φ1 between the Cp ring and the (thio)acetal.
The influence of the torsion angle of the thioacetal or acetal on the orientation of the
diphenylphosphine group is unclear, as is the influence of substitution of both the Cp and
the Cp in Connell’s case as compared to compound 25, which is only substituted on one Cp
ring. To compare the orientation of the phosphine, two angles are defined, φ2 as the dihe-
dral angle C(17)–P(1)–C(1)–C(5) and φ3 as the dihedral angle C(11)–P(1)–C(1)–C(5) and
the corresponding angles in Connell’s diacetal. In compound 25 φ2 has a value of 87.1(2)° in
comparison to the φ2 in Connell’s diacetal measuring 107.4° and 115.7°, for φ3 the value is
27

48. 2 Synthetic approaches towards PSiP-Pigiphos
170.0(2)° in 25 and 147.8°, 138.9°, respectively, in Connell’s diacetal (cf. Table 3). φ2 and φ3
can be used as an indicator for the orientation of the phosphorus’ lone-pair.
φ2 φ3
Compound 25 [°] Compound 27 [°] Compound 25 [°] Compound 27 [°]
87.1(2) 107.4 170.0(2) 147.8
115.7 138.9
Table 3: Dihedral angles along the phosphine–ferrocene bond.
The phosphorus lone-pair, appears to be oriented towards the thioacetal moiety in 25. At the
same time the value of φ1 implies an orientation of the acidic hydrogen of the thioacetal to-
wards the phosphine lone-pair. As the measured distance between C(23) and P(1) of 3.42 Å
is comparable to the sum of the van der Waals-radii of phosphorus and carbon, which would
be 3.5 Å,[174,175]
the influence of hydrogen bonding between the phosphorus lone pair and
the acidic proton at C(23) should be taken into consideration. Such an interaction would also
explain the coupling constant observed in 1
H-NMR of JPH = 4.5 Hz. Interestingly the chemical
shift of the acidic proton on C(23) at δ 5.26 ppm, is shifted downfield in comparison to the
chemical shift of the corresponding proton in 28 at δ 4.87 ppm. This implies, that although
the hydrogen bonding between P(1) and the acidic proton on C(23) might lead to a higher
electron density at the hydrogen, the weakening of the bond between the proton and C(23),
due to the interaction with the posphine, is strong enough to leave the proton more exposed
to the magnetic field causing the downfield shift. Another explanation could be the differing
position of that particular proton towards the ferrocene and therefore its exposure to the fer-
rocene’s ring current. As, in solution, dynamic behaviour of the thioacetal is expected, one
could expect this influence to be neglect-able, unless the thioacetal position is fixed due to
the suggested hydrogen bond interaction. This type of interaction may facilitate the already
facile lithiation of the thioacetal by stabilising the lithiated species due to interactions between
the lithium cation and the phosphine lone-pair. Still, the high sterical demand at the reactive
centre may cause difficulties in an attempt to couple two ferrocenyl thioacetals over a silane
moiety, as has been demonstrated by further experiments (vide infra).
Due to the low overall yield of the synthesis of the thioacetal 25, experiments that would allow
the synthesis of a PSiP-Pigiphos derivative were first tested using ferrocenyl dithiane 29. Two
equivalents of lithiated ferrocenyl dithiane 29 were allowed to react with one equivalent of
dichlorodimethylsilane yielding the silylated dithiane 28 and starting material in a 1:1 ratio.
As the reaction might occur in an SN 2 like fashion, this is not surprising, since a second substi-
28

49. 2.3 Synthetic Results
tution of the silicon moiety would be hindered due to the sterically very demanding transition
state necessary for the formation of the desired product (cf. Scheme 34). In order to force a
Fe
Si
S S
Cl
Fe
SS
‡
Fe
H
SS
2
1. 2 equiv n-BuLi
2. 1 equiv SiMe2Cl2
Fe
Si
SS
Fe
H
SS
Cl
+
sterically hindered transition state
29 28 29
Scheme 34: Reaction of dithiane 29 n- BuLi and dichlorodimethylsilane and the suggested sterically
hindered transition state for the SN 2-reaction of chlorosilane 28 with dithiane 29.
second substitution of the silyl dithiane 28, several experiments were performed. Cleavage of
the dithiane to give the carbonyl was attempted to reduce the steric crowding around silicon.
Although compound 28 showed high stability and was even stable in contact with water, the
conditions used for thioacetal cleaveage[176]
would most probably lead to a reaction at the
resulting highly electrophilic silicon centre. Nonetheless, attempts to deprotect the intermedi-
ate were performed using the milder conditions reported by Soderquist et al..[177]
Even under
these conditions, decomposition of the deprotected product was observed.
Aside from the thioacetal cleavage, forcing an SN 1 type mechanism may be another option
to induce the second substitution at the silicon moiety. Because of the high Lewis acidity of
silicon, if a halogen scavenger can generate an even slight concentration of a corresponding
silylenium, even the sterically demanding lithiated dithiane 29 could react. However, reaction
with silver bis(triflimide) led to oxidation of the ferrocene moiety to ferrocenium characterised
by the deep blue colour of the resulting reaction mixture. The 1,3-propanedithiol moiety also
appeared to have been cleaved under these conditions, as the resulting reaction mixture pos-
29

50. 2 Synthetic approaches towards PSiP-Pigiphos
sessed a strong garlicky stench. Using sodium BArF as halogen scavenger lead to a mixture
of several compounds, which could not be completely separated. 29
Si-1
H-HMBC-NMR showed
two silicon species within a small fraction separated by silica flash column chromatography.
Neither of these could by isolated or fully characterised. The other fractions collected con-
tained a diverse mixture of compounds, resulting from decomposition of the starting material.
Due to the difficulties described and the foreseeable increase of difficulties for compound 25
resulting from the even higher steric demand of the ortho diphenyl phosphine group, further
attempts towards the double substitution of the silicon moiety were abandoned in favour of
further study of silyl substituted derivatives of 25 (vide infra), and the investigation of PSiP-
pincer ligands (cf. chapter 3).
2.3.4 Synthesis and characterisation of PSi derivatives of dithiane 26
Since double substitution of the silicon moiety was unsuccessful, the remaining phos-
phino dithiane 25 synthesised was used for preliminary studies of bidentate PSi ligands. The
chlorodimethylsilyl derivative 30, as well as the dimethylsilyl derivative 2, were produced by
quenching lithiated dithiane 25 with the appropriate chlorosilane (cf. Scheme 35).
Fe
H
S S
PPh2
n-BuLi
SiCl2Me2
n-BuLi
SiHClMe2Fe
Si
Cl
S S
PPh2
Fe
Si
H
S S
PPh2
<16.8 % 82 %
30 25 2
Scheme 35: Synthesis of the PSi ligands 30 and 2
While compound 25 could not be fully converted to compound 30 and the resulting yield
turned out to be quite low, the synthesis of silane 2 worked well and in a satisfactory
yield. Both compounds showed high stability in air and even against moisture, which is
particularly surprising in the case of the chlorosilane 30. It seems that the steric hindrance
to the accessibility of the silicon moiety paired with the relatively high electron density at
the silicon centre reduces its Lewis acidity to a point at which it is almost inert towards
water. The proximity of the phosphine to the silicon moiety seems to cause an interaction
between the two hetero atoms. This can be seen by 29
Si-NMR in which a significant upfield
30

51. 2.3 Synthetic Results
shift of ∆δ –14.0 ppm is observed, when comparing the chlorosilane 28 (δ 23.9 ppm) and
the phosphine bearing derivative 30 (δ 9.9 ppm) to each other. In case of the silane 2, a
coupling between the P-atom and one of the methyl groups bound to the silicon is observed
in 31
P,1
H-HMBC-NMR, providing evidence of a phosphine–silane interaction.
Out of a number of complexation experiments, only treatment of the silane 2 with [Pt(PPh3)4]
in C6D6 at rt led to an identifiable product formed by Si–H activation. This process seems
to occur similarly to the Si–H activation of the (ortho-phosphinophenyl)silane 31 with
Pt(0) as reported by Takaya et al.[178]
Takaya et al. report the formation of a complex
with trigonal-bipyramidal geometry, with an additional PPh3 coordinated to the platinum
moiety as well as the hydride and the PSiP-ligand. This complex seems to be formed via a
square-pyramidal intermediate, which is observable during the first 4 h of the reaction but is
subsequently condumed (cf. Scheme 36).
P
Si
P PPh2H MePh2
Pt(PPh3)4
3 PPh3
Pt
PPh3P
H Si
P
Ph2
Me
Ph2
isomerisation
rt
PtH Si
P
PPh3 Me
31
Scheme 36: Formation mechanism of the PSiP-platinum complex of 31 reported by Takaya et al.[178]
(the phenyl groups at the PSiP ligand in the product are omitted for simplicity)
In case of the formation of the Pt-2-complex, no intermediate was observed by NMR, of
course the fact that 2 is a bidentate ligand should facilitate the coordination and reaction
at the metal centre. The product would, therefore, form by coordination of the phosphine
moiety to Pt(0), followed by formation of an η2
bond with the sigma orbital of the Si–H bond,
which then leads to Si–H activation, resulting in a square planar cis-Pt(II)-2-complex with the
coordination site trans to the silyl-donor occupied by triphenylphosphine (cf. Scheme 37).
Structural hypotheses are based solely on NMR experiments. In 31
P{1
H}-NMR two phosphorus
signals, with the same intensity, showing platinum satellites were observed among signals for
free triphenylphosphine, 2 and Ph3PO. Together with the observation of a hydride signal in
the 1
H-NMR spectrum, this suggests a square planar Pt(II) species. The small 31
P–31
P coupling
constant (JPP = 15.7 Hz) is an indication of the cis orientation of the two phosphine ligands.
A further indicator of cis configuration is the magnitude of the Pt–P coupling constant, which
31

52. 2 Synthetic approaches towards PSiP-Pigiphos
Fe
Si
SS
HP
PtLn
Fe
Si
SS
HP Pt
Ln
Fe
Si
SS
HP Pt
PPh3
L = PPh3
Scheme 37: Suggested formation path of Pt(II)-2-complex.
has been shown to decrease with increasing trans influence of the adjacent ligand.[179]
The
coupling constants of the two phosphines in the Pt(II)-2 complex differ significantly and are
comparable to the data reported by Chan et al.[180]
Chan et al. report coupling values for
the phosphine trans to the hydride between 2512 – 2716 Hz, while coupling constants for the
phosphine trans to the silyl donor are much lower at 1280 – 2055 Hz.
5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 -0.5 -1.0 -1.5 -2.0 -2.5 -3.0 ppm
0
5
10
15
20
25
30
35
ppm
PtHSi(CH3
)2
CHCp
CHCp
PPh2
on 2
Ph3
PO
coordinated
PPh3
Figure 3: 1
H–31
P-HMBC spectrum (delay set for J = 8 Hz) of Pt(II)-2 complex.
32

53. 2.3 Synthetic Results
Therefore, it seems that the 31
P-NMR signal at δ 30.4 ppm with JPtP = 1560 Hz corresponds to
the phosphine trans to the silyl group, while the signal at δ 14.1 ppm with a coupling constant
JPtP = 2487 Hz corresponds to the phosphine trans to the hydride. The signal at δ 14.1 ppm
can thus be assigned to the phosphine in ligand 2. This was confirmed by 1
H–31
P-HMBC,
which shows a correlation between that phosphorous centre and the ferrocene protons (cf.
Figure 3). There is also a clear correlation observed between the methyl groups on the silicon
and the triphenylphosphine, as well as between the hydride and the two phosphines.
29
Si–195
Pt coupling extracted from the 29
Si-INEPT-NMR spectrum has a value of JSiPt = 1114 Hz,
which is comparable to values for similar complexes found in literature.[138]
In the 1
H-NMR,
a coupling to 195
Pt was found for the hydride, as well as the methyl groups on the silicon
(JPtH = 1065 and 40 Hz, respectively). In order to measure the 195
Pt-NMR shift, a 1
H–195
Pt-
HMQC was run with a delay adjusted to the coupling of the hydride to the platinum of
JPtH = 1065 Hz. The platinum shift was found to be at δ –5235 ppm showing correlation to
the hydride and the methyl groups on silicon (cf. Figure 4).
-0.50.00.51.01.52.0 -1.5 -2.0 -2.5 -3.0 -3.5 -4.0 ppm
-5400
-5380
-5360
-5340
-5320
-5300
-5280
-5260
-5240
-5220
-5200
-5180
-5160
-5140
-5120
-5100
-5080
-5060
-1.0
ppm
A
A
A
A
A
A
A
A
B B
B B
B B
B B
A: 195
Pt–hydride cross-peaks B: 195
Pt–Si(CH3
)2
cross-peaks
Figure 4: 1
H–195
Pt-HMQC spectrum (delay set fot J = 1065 Hz) of Pt(II)-2 complex.
33

54. 2 Synthetic approaches towards PSiP-Pigiphos
2.4 Summary
The synthesis of a PSiP-Pigiphos ligand 1 was investigated following two general approaches.
Although the synthesis of the tridentate ligand was unsuccessful, a synthetic route to an
asymmetric ferrocenyl PSi-ligand 2 was established. This ligand underwent Si–H activation
with [Pt(PPh3)4] to form a square-planar hydrido-triphenylphosphino-2-platinum(II) com-
plex. This complex is of interest for further investigations concerning its catalytic activity
as well as ligand exchange mechanisms. Due to the failure of the attempted PSiP-Pigiphos 1
synthesis, an approach towards a different kind of PSiP-pincer became a matter of interest (cf.
Chapter 3).
34

55. 3 Synthetic approaches towards a chiral PSiP-Pincer ligand
3.1 Introduction
Since major difficulties were encountered in the synthesis of a PSiP-Pigiphos ligand (cf. chapter
2), a simpler molecular structure became a matter of interest. Because the bulk of the prob-
lems were primarily related to the carbon spacer between the ferrocene and the silicon moiety,
exclusion of the spacer resulting from direct silylation of the ferrocene moieties should there-
fore alleviate the problems encountered in the PSiP-Pigiphos synthesis. This would result in a
PSiP pincer-like ligand, that would form a five membered metallacycle upon Si–H activation.
Such a ligand should fullfill the requirements that were already set out for the PSiP-Pigiphos
and therefore, represent the first chiral PSiP-pincer ligand.
3.1.1 Pincer ligands
Ever since the first synthesis of a pincer type ligand by Moulton and Shaw,[181]
this platform
has been of great interest. Whereas pincer complexes of the ECE-type (cf. Scheme 38) bearing
a central aryl ring, which interacts with the metal centre via its anionic Cipso atom,[181–186]
were
of interest during the first twenty years of pincer ligand chemistry, today a much larger variety
of pincer ligands are known. The great variety of pincer ligand systems is due to diversity of
applicable ligand backbones.[185,187–189]
E
E
E
E
M X
E = NR2, PR2, SR
M = Ni, Pd, Pt, Rh, Ir, Sn
Scheme 38: Generalised structure of ECE-pincer ligands and their complexes as first reported by Shaw
and co-workers[181,182]
and van Koten et al.[183,184]
The pincer ligand platform has several defining characteristics. Pincer ligands are tridentate
ligands, which form κ3
complexes around a metal centre and contain two metallacycles. They
bear two lateral donor atoms and a central carbon that forms an ipso-carbon-metal bond upon
complex formation, usually through C–H activation.[190]
As a consequence of these features,
35

56. 3 Synthetic approaches towards a chiral PSiP-Pincer ligand
the resulting pincer complexes are highly stable. It has been reported by Shaw[191,192]
that the
introduction of two five-membered metallacycles also increases the thermodynamic stability
of these systems. It is the high thermal stability paired with the high reactivity, that arises
from the strong σ-donor effect of the ipso-carbon, which make pincer complexes attractive
for use in catalysis.[193]
Pincer complexes have shown a variety of applications not only in
catalysis,[188,194–197]
but also as chemical sensors and chemical switches.[187]
3.1.2 Pincer-like PSiP-ligands
Today the term pincer-like ligand is often used to designate ligands with similar features
as actual pincer ligands. They include tridentate complexes with carbene centres or even
nitrogen instead of carbanions.[198]
Among these alternative ‘pincer’ ligands/complexes the
PSiP-pincer like ligands are probably the closest example to the original pincer ligands.
Although transition metal-silicon chemistry is well-known[125][199]
only a few examples of
silyl donors in a framework of ancillary ligands have been reported.[133,134,200]
The first
syntheses of pincer-like NSiN-ligands and their complexes have been performed by Tilley and
co-workers,[142–144]
while the Turculet group has claimed the first synthesis of a pincer-like
PSiP ligand.[132]
Since then, there have been a remarkable number of publications concerning
the complexes of this PSiP pincer-like ligand and their chemical properties as well as their
catalytic use.[132,140,201]
The ready accessibility of coordinatively unsaturated metal com-
plexes[202]
or even electron deficient late transition metal complexes[140]
of the PSiP-ligand
(cf. Scheme 39) is a direct consequence of the strong trans influence of the silyl donors (cf.
Chapter 2.1.2) introduced into the pincer framework.
Cy2P
SiMe
Cy2P
Ru N
SiMe3
SiMe3
Si
PCy2
PCy2
MMe
H
Cl
M = Rh, Ir
Scheme 39: Coordinatively unsaturated and electron deficient pincer-like PSiP complexes reported by
Turculet and co-workers.[140,202]
36

57. 3.2 Synthetic strategy
3.1.3 Aim of the project
The synthesis of an asymmetric pincer-like ferrocenyl based PSiP ligand 3 is the main goal
of this project. A secondary objective of the project was the synthesis of a structurally
analogous PPP ligand 4 to allow for comparative studies of the PSiP pincer-like ligand as well
as PigiPhos (cf. Scheme 40). Because of the five membered metallacycles, which are formed
by complexation, the central phosphorus donor atom is expected to be closer to the metal
centre. This should lead to a distinctive trans influence and, therefore, the resulting Ni(II)-PPP
complex should show comparable reactivity to the PSiP-Pigiphos analogue 32 described in
Chapter 2.1.3. Although the synthesis of 4 has been reported by Butler,[203]
only the racemate
was isolated and no complex chemistry has been done with this type of ligand to date.
Fe
Fe
Si
P P
H
R
Ph2Ph2
Fe
Fe
P
P P
Ph
Ph2Ph2
R = Me, Ph
3 4
Scheme 40: Proposed asymmetric pincer-like PSiP ligand 3 and its PPP analogue 4.
3.2 Synthetic strategy
In order to introduce planar chirality at the ferrocene moieties during the synthesis of the
PSiP-pincer like ligand 3 the sulfoxide route described by Kagan and co-workers[75]
was chosen
(cf. Paragraph 1.1.2.2). Starting from ferrocene the chiral p-tolyl-ferrocenyl-sulfoxide 16 is
easily synthesised as reported by Ribière et al.[204]
Selective ortho-lithiation then should yield
either the phosphine 33 or the silane 34 as needed. In a second step, the sulfoxide can be
substituted by another electrophile using t-BuLi (cf. Scheme 41).
An analogous route should yield the corresponding PPP analogues 4. As double lithiation of a
37

58. 3 Synthetic approaches towards a chiral PSiP-Pincer ligand
Fe
Fe
Si
P P
H
R
Ph2Ph2
Fe
S p-tol
O
1. LDA
2. ClPPh2
3. BH3-THF
Fe
S p-tol
OPPh2
H3B
Fe
S p-tol
O
1. LDA
2. RHSiCl2 Fe
Fe
Si
H
S S
toltol
O
O
R
1. t-BuLi
2. RHSiCl2
3. NEt3
1. t-BuLi
2. RHSiCl2
Fe
Fe
Si
H
P P
Ph2Ph2
R
R = Me, Ph
34
35
3
3
Scheme 41: Proposed synthetic routes towards the PSiP-pincer like ligand 3.
molecule would be necessary in order to obtain (SFc,SFc)-3 or (RFc,RFc)-4 the convergent route
yielding the opposite enantiomers should be more feasible.
3.3 Synthetic challenges
3.3.1 Synthetic approach towards the PSiP-pincer like ligand 3
The synthetic approach towards the borane protected phosphine 33 as described by Riant
et al.[75]
was reproduced without any problems in a reasonable yield. Problems were not
faced until the attempt of double substitution of dichloromethylsilane in the second step,
which yielded none of the desired material. Reaction of the deprotected phosphine sulfoxide
also failed to yield the desired product upon lithiation with t-BuLi and subsequent quenching
with dichloromethyl silane. Approaching the target compound by first double substituting the
silicon moiety gave the bis sulfoxosilane 34 in a 30 % yield. While the second substitution gave
only minimal amounts of what could be considered to be the target compound, considering
31
P-NMR (δ –18.5 and –19.71) and MALDI-MS (m/z calcd: 783.12 found: 784.12 [M + H+
]).
Several attempts of optimising reaction conditions (temperature, solvents, reaction time) were
unsuccessful, despite the scale of the reaction, only amounts suitable for NMR analysis could
be isolated. Therefore other approaches had to be taken into consideration (cf. Chapter 3.3.3).
38

59. 3.3 Synthetic challenges
3.3.2 Synthetic approach towards the PPP-pincer analogue 4
Similarly to the synthesis of the PSiP-pincer like ligand, 3, two approaches were fol-
lowed. In the first approach, the ferrocenyl sulfoxide 16 was first substituted by using
chlorodiphenylphosphine and subsequently protected with borane. The resulting phos-
phinoferrocenyl sulfoxide 33 was then reacted with one equivalent of t-BuLi and half an
equivalent of dichlorophenylphosphine. As was observed during the synthesis of 3 (vide
supra), this approach failed to yield the desired product. Coupling the ferrocenes over the
phenylphosphine moiety in the first reaction step, gave a low yield of about 18 % in inital
efforts. The second step led to only trace amounts of the target material in a product mixture.
The product was identified in the mixture by ESI-HRMS (calcd: 846.1198, found: 846.1254
[M+
]), encouraging further effort in the improvement of the synthesis. Stepwise lithiation of
the bissulfoxophosphine 5 was attempted in order to avoid a route over a double anion (cf.
Scheme 42).
Fe
Fe
P
P
Ph
Ph2
Fe
Fe
P
S S
toltol
O
O
Ph
1. t-BuLi
2. ClPPh2 S
tol
O
1. t-BuLi
2. ClPPh2 Fe
Fe
P
P
Ph
Ph2
P
Ph2
5 4
Scheme 42: Proposed stepwise lithiation of 5.
Initial efforts to develope a one-pot reaction lead to the formation of a mixture of phosphines.
Introduction of a work-up and filtration over silica in DCM after the first lithiation increased
the yield of the desired final product to an NMR-detectable amount. Seperation by flash
column chromatography gave a mixture of three major compounds, as shown by HPLC
(OD-H, n-hexane/i-PrOH 95:5, 0.7 ml/min, tR: 8.05, 8.47, 8.72 min). As further attempts at
purification were unsuccessful, preparative HPLC was used to further separate the mixtures
under the optimised conditions determined by analytical HPLC (vide supra). These efforts led
to a slightly better but still incomplete seperation. Three fractions were collected, of which
the second (tR: 7.88 – 8.49 min) contained the majority of the desired product, which was
fully characterised. One of the side products separated, could also be characterised and was
found to be diphosphine 35 (cf. Scheme 43).
39

60. 3 Synthetic approaches towards a chiral PSiP-Pincer ligand
Fe
Fe
P
P
Ph
Ph2
P
Ph2 Fe
Fe
P
P
Ph
Ph2
4 36
Scheme 43: Products characterised after preperatory HPLC separation.
3.3.3 Explanation for the synthetic difficulties
A closer look at the synthetic difficulties encountered during the synthesis of the PSiP pincer-
like ligand 3 and its PPP analogue 4 reveals that particular reaction steps turn out to cause
major difficulties. Firstly any reaction involving the coupling of the ferrocenes via a central
moiety gives low yields. If the coupling is carried out in the same step as the cleavage of the
sulfoxide no desired product could be isolated (cf. Table 4).
As in the case of coupling over a phosphine moiety, various 31
P-NMR signals corresponding to
phosphine oxides were found, the conclusion seemed preeminent, that an oxide transfer from
the sulfoxide to the electrophile takes place. Interestingly, the free t-butyl-p-tolyl sulfoxide
formed during the sulfoxide cleavage seems to perform the oxidation more efficiently than the
bound sulfoxide leading to no product. Therefore it can be assumed, that t-butyl-p-tolyl sul-
foxide might also play a role in the substitution reaction forming the bisferrocenyl species 34
and 5, thus lowering the yields further. These observations serve to emphasise that a direct
synthesis of the PSiP pincer-like 3 and its PPP analogue 4 from the sulfoxide precursors could
not be achieved satisfactorily. As a consequence the synthesis of an inert building block was
investigated (cf. Chapter 5.3).
3.4 The sulfoxophosphine ligand 5
3.4.1 Structure discussion
Over the course of the synthetic route towards the PPP ligand 4 the SPS compound 5 was syn-
thesised as an intermediate in yields up to 45 %. Compound 5 showed interesting features in
40

61. 3.4 The sulfoxophosphine ligand 5
Fe
Fe
R''
E
Fe
Fe
R''
E
P P
Ph
PhPhPh
S S
toltol
O
O
LiR'
ECl2R''Fe
R
S
tol
O
or
E = SiH, P
R = PPh2, H
R' = t-Bu, i-Pr2N
R'' = Me, Ph
E R R R yield [%]
Si–H Me PPh2 t-Bu 0
Si–H Me H i-Pr2N 16 – 33
P Ph PPh2 t-Bu 0
P Ph H i-Pr2N 18 – 45
Table 4: Generalised scheme for the coupling step in the synthesis of 3 and 4 and the yields corre-
sponding to the respective reactions.
1
H-NMR. As the epimerisation barrier for phosphines usually lies around 30 kcal/mol[205,206]
5
may be described most strictly as a C1 symmetric molecule at rt, therefore the hydrogen atoms
corresponding to each other on the ferrocenyl and tolyl groups are diastereotopic, hence the
different chemical shifts. The large difference in the chemical shifts of the two Cp rings with
a ∆δ of 0.72 ppm is remarkable. Although this observation seemed quite astonishing at first,
X-ray structure determination of crystals grown from DCM/n-hexane gave rise to a possible
explanation for this strong shift (cf. Figure 5).
The tolyl group on S(1) is oriented in such a way, that the aryl ring lies 3.49 Å away from
the next Cp carbon bound to Fe(1), facing the Cp with the ring plane of the tolyl group.
This T-shaped orientation to each other may be due to a π–π interaction. Therefore, it can
be assumed that a similar conformation is predominantly present in solution and the ring
current of the tolyl group enhances the field at the Cp hydrogens leading to the upfield
shift of about ∆δ 0.72 ppm observed in 1
H-NMR. The aryl ring on S(2) is oriented in nearly
the opposite direction with regards to the ferrocene moiety (cf. Table 5), comparable to the
reported structure of p-tolylferrocenyl sulfoxide.[207]
As a consequence the S(2)-aryl lies face
to face with the phenyl ring on P(1) in an almost parallel fashion (angle between ring planes
is 5.87°) at a distance around 3.8 Å, implying that a parallel-displaced π–π-interaction is
present.[208]
41

62. 3 Synthetic approaches towards a chiral PSiP-Pincer ligand
P1 C2
C1
S1
O1 C11
C35
C19
C18
O2
S2
C28
Fe2
Fe1
Figure 5: X-ray crystal structure of the SPS-ligand 5.
dihedral Angle [°] dihedral Angle [°]
C(11)–S(1)–C(1)–C(2) –115.8(2) C(28)–S(2)–C(18)–C(19) 89.8(2)
O(1)–S(1)–C(1)–C(2) 132.67(19) O(2)–S(2)–C(18)–C(19) –21.0(2)
C(35)–P(1)–C(2)–C(1) 163.07(19) C(35)–P(1)–C(19)–C(18) –75.3(2)
Table 5: Selection of dihedral angles of compound 5.
The S(1)tolyl facing the Cp on Fe(1) also has a significant impact on the C(1)–S(1)–C(11)
angle which is widened by about 5° in comparison to the C(18)–S(2)–C(28) angle, while the
O–S–Fc angle is conversely widened around the S(2) moiety (cf. Table 6).
Angle [°] Angle [°]
C(1)–S(1)–C(11) 104.24(10) C(18)–S(2)–C(28) 99.42
O(1)–S(1)–C(1) 104.93(11) O(2)–S(2)–C(18) 109.05(10)
O(1)–S(1)–C(11) 106.28(10) O(2)–S(2)–C(28) 106.11(11)
Table 6: Bond angles around the sulphur atoms of 5.
Although one would expect that bond angles around the phosphorus atom should increase
42

63. 3.4 The sulfoxophosphine ligand 5
with more sterically demanding groups,[209]
compound 5 shows narrower angles around P(1)
than diferrocenylphenylphosphine reported by Houlton et al.[210]
except for the angle between
a ferrocenyl substituent and the phenyl substituent (cf. Table 7).
compound 5 diferrocenylphenylphosphine[210]
Angle [°] Angle [°]
C(2)–P(1)–C(19) 98.55(10) C(24)–P–C(35) 100.0(5)
C(2)–P(1)–C(35) 99.30(10) C(35)–P–C(51) 101.0(5)
C(19)–P(1)–C(35) 99.30(10) C(24)–P–C(51) 98.6(5)
Table 7: Comparison of bond angles around the phosphine of compound 5 and diferro-
cenylphenylphosphine.
3.4.2 Complexation Experiments
A variety of experiments were performed in order to prepare complexes of ligand 5. The
focus of the complexation experiements was limited to d8
metals, except for one complexation
experiment using a Pd0
precursor. Most of the complexation reaction products could not be
completely characterised, as only inconclusive 1
H-NMR spectra were obtained particularly
in case of the rhodium and iridium complexes. Therefore 31
P-NMR and HRMS were used as
indicators for complexation when possible. Evidence of complex formation was found in five
experiments, for which a variety of MS and NMR methods were used (cf. Table 8).
Metal precursor MS 31
P-NMR 1
H-NMR NOESY X-ray
[(C2H4)2RhCl]2 n.a. × inconclusive n.a. n.a.
[(COD)RhCl]2 × × inconclusive n.a. n.a.
[(COE)2IrCl]2 × × inconclusive n.a. n.a.
[Pd(COD)Cl2] × × × n.a. ×
[Pt(COD)Cl2] × × × × n.a.
[Pd2(dba)3] n.a. inconclusive n.a. n.a. n.a.
Table 8: Analytical data available for 5-metal complexes.
Although analytical data for the rhodium complexes is sparse, formation of a complex is
43

64. 3 Synthetic approaches towards a chiral PSiP-Pincer ligand
clearly indicated by 31
P-NMR. A doublet signal at δ 56.5 ppm with a coupling constant JRhP of
170 Hz independent of the precursor used is observed. In order to rule out the possibility that
the doublet observed corresponds to two signals from two different species, the spectrum was
measured using a 300 MHz, 400 Mhz and 500 MHz NMR confirming the coupling. (cf. Figure
6)
54.0 ppm55.056.057.058.0
202.1 MHz 31P{1H} NMR
162.0 MHz 31P{1H} NMR
121.5 MHz 31P NMR
Figure 6: 103
Rh –31
P-coupling measured at different field strengths.
In addition to the signal showing the Rh–P-coupling, a singlet at δ 26.4 ppm was observed in
case of the rhodium complex synthesised using [Rh(C2H4)2Cl]2 as a precursor. This implies
that a portion of the ligand 5 is only bound over the sulfoxide moieties to rhodium thereby
leaving the phosphorus uncoordinated. If ligand 5 is reacted with [Rh(COD)Cl]2 the singlet
signal is shifted upfield by ∆δ 0.7 ppm. This suggests that some of the precursor’s ligands
might be included in the complex formed, causing differences in chemical shifts. From this
information, the presence of a large variety of complexes from mononuclear to multi nuclear
complex clusters could be possible. MALDI-TOF-HRMS measurements show a single signal
at m/z 859.9564 corresponding to the molecular formula C40H35Fe2O2PRhS2 which would
fit the formula [RhL]+
ion, with L being 5. In case of the reaction of 5 with [Ir(COE)2Cl]2
a brown powder was isolated that showed a 31
P-NMR shift of δ –4 ppm and a weak signal
44

65. 3.4 The sulfoxophosphine ligand 5
in MALDI-TOF-HRMS for C80H70Fe4IrO4P2S4, corresponding to [IrL2]+
. These were the
only indications of possible complex formation with iridium(I) and ligand 5. Reacting
[Pd(COD)Cl2] with ligand 5 yielded a red powder that showed a 31
P-NMR shift at δ 44 ppm.
(In the 1
H-NMR significant chemical shifts for the protons close to the metal moiety could be
observed as well as line broadening. Some signals are more strongly affected than others in
terms of those two parameters.) MALDI-TOF-HRMS showed a weak signal at 859.9561, which
corresponds to [PdL]+
. Attempts to synthesis a Pd(0) complex by reacting [Pd(dba)] with
ligand 5 gave inconclusive results in 31
P{1
H}-NMR, which showed two very broad signals at
δ 28.3 and 26.6 ppm. A platinum(II) complex of 5 was synthesised by dissolving the ligand
in DCM with [Pt(COD)Cl2] yielding a yellow powder. 31
P{1
H}-NMR of the compound showed
a peak at δ 21.1 ppm with platinum satellites having a coupling constant of JPPt = 3.7 kHz,
while in MALDI-TOF-HRMS a signal corresponding to [PtLCl2 + Na]+ was detected. All of
the complexes mentioned showed poor solubility in ether, toluene or benzene, but they were
moderately soluble in chloroform and THF. They also showed moderate to good solubility
in DCM and pyridine. In order to obtain single crystals, a variety of crystallisation methods
were applied, using different solvent systems. While most attempts resulted in decomposition
of the complexes in solution or precipitation of a powder, crystallisation of [PdLCl2] by gas
phase diffusion of benzene into a THF solution of the complex at –20 ◦
C was successful. The
resulting single crystals were of poor quality, preventing refinement of the crystal structure
further than to an R-value of 8.41 %. The low quality of the crystals may be explained by
the high solvent to complex ratio in the crystals and the low crystallisation temperature. Two
molecules of the platinum complex crystallised together with ten benzene molecules and one
THF molecule. In addition to the disorder of the solvent molecules, this might also have lead
to cracks in the crystals due to solvent evaporation during the short period of time the crystals
were at rt. However, a reasonable structural model could be obtained from the solution of the
crystal structure, showing that the phosphine as well as one sulfur moiety coordinate to the
platinum centre, while the other two coordination sites are occupied by two chlorido ligands
(cf. Figure 7).
This structure was also corroborated by 2D-NMR data of the platinum(II) complex. The
NOESY spectrum showed a contact between protons of the tolyl group on S(2) (δ 8.06 ppm)
and protons on the phenyl ring (δ 7.70 ppm; cf. Scheme 44). This contact is only possible if
one sulfoxide is not coordinated to the metal centre, thus allowing it to move freely into a
conformation allowing contact.
The fact that only one sulfur is bound to the metal centre is not surprising, as it is known
45

66. 3 Synthetic approaches towards a chiral PSiP-Pincer ligand
P1
S1
S2 O1
O2
Pd1
Cl2
Cl1
Figure 7: Structure of the dichloropalladium(II) complex of 5.
Fe
Pt SCl
O
P
Fe
SO
Cl
NOE contact
5
Scheme 44: Observed NOE contact in the dichloroplatinum(II) complex of 5.
that sulfoxide bound over the sulfur moiety to the metal have a strong trans labilising
effect.[211–213]
Therefore, the only accessible position for the second sulfur moiety, which
lies trans to the other sulfur donor, in case of a square planar coordination around the
metal, is strongly disfavoured. Only coordination by the oxygen atom can be taken into
consideration,[214]
but seems unlikely considering the arguments mentioned above. Tem-
perature dependent 1
H-NMR of the palladium complex showed, that at 303 K the signals
corresponding to both tolyl groups are close to the fast exchange limit with regard to rotation
around the S–tol bond. Also the phenyl group shows fast exchange with respect to rotation
around the P–Ph bond. When the temperature is decreased, the signals corresponding to the
phenyl protons and the protons on the tolyl group on S(1) (cf. Figure 7) become broader,
46

67. 3.4 The sulfoxophosphine ligand 5
with the phenyl proton signals coalescing at 223 K, while the protons on the tolyl group on
S(2) remain in fast exchange (cf. Figure 8). Assuming that the chemical shift difference
between the exchanging protons on the corresponding aromatic rings are similar, when
the slow exchange limit is reached, the energy barrier to the rotation is the lowest in case
of the tolyl group on S(2). This implies that the coordination of S(1) and the phospho-
rus to the palladium have an effect on the rotational barrier of the attached aromatic systems.
7.07.58.08.5
223
233
243
253
263
273
283
293
303
ppm
T/KS(1) P(1) S(2) P(1) S(1) S(2)
Figure 8: Temperature dependent 1
H-NMR of the dichloropaladium(II) complex of 5.
3.4.3 Catalytic experiments
The complexes of ligand 5 were tested for their catalytic activity in selected reactions. The
rhodium complex of ligand 5, synthesised from [Rh(COD)Cl]2, was tested in Miyaura-Hayashi
reaction following the reaction procedure by Dornan et al.[215]
Initial attempts at 40 ◦
C gave
high yields (up to 99 %), but negligible enantiomeric excess. At 0 ◦
C no product is observed,
however, the best conditions were found to be around room temperature giving 60 % yield
and up to 19 % ee (cf. Table 9). Although the enantiomeric excess achieved is only marginal,
47

68. 3 Synthetic approaches towards a chiral PSiP-Pincer ligand
it implies that the chiral ligand is somehow involved in the catalytic cycle.
O B
OHHO
cat.
EtOAc
O
Rha
, PhB(OH)2, T, t, yield ee
mol% equiv ◦
C h % %
4 4 40 24 99.5 rac
4 4 40 1 40 rac
4 1.5 40 24 95 rac
2 4 40 1 60 rac
4 4 0 24 0 n.a.
4 4 rt 24 14 16
4 4 rt 4 11 15
a
under the assumption that the rhodium complex
of ligand 5 has a molecular formula of the type
[C40H35ClFe2O2PRhS2]n
Table 9: Rhodium catalysed Miyaura-Hayashi reaction.[215]
The dichloropalladium complex of 5 was tested in allylic substitution of diphenylallyl acetate
with dimethylmalonate. The isolated complex was used initially, and was synthesised starting
from [Pd(COD)Cl2]. The reaction was performed at 0 ◦
C for 4 h yielding 7 % of the desired
product with 82 % ee of the (S) enantiomer. As the isolated dichloropalladium complex
showed such low activity, a catalysis was run with the palladium catalyst generated in situ,
by adding ligand 5 and bis((1,3-diphenlallyl)bromopalladium(II)). Although a higher yield
was achieved (90 %), the main catalytically active species in the reaction mixture seemed
to be the precursor itself, as only a racemic mixture (2 % ee) was isolated. The problems
were overcome by using bis(allylchloropalladium(II)) as precursor to generate the catalyst
in situ with ligand 5. In a first attempt a yield of 97 % at an enantiomeric excess of 77 % ee
was achieved. Change of parameters such as solvent, temperature and base did not show
significant impact on the enantiomeric excess (cf. Table 10). Still further optimisation may be
considered.
48

69. 3.5 Summary
Ph Ph
OAc O
MeO
O
OMe
Pd cat.
N,O–bis(trimethylsilyl)
acetamide Ph Ph
O
OMe
O
MeO
Pd, additive T, t, solvent yielda
ee
mol% ◦
C h % %
5 LiOAc rt 16 AcN 97 77
5 NaOAc rt 16 AcN 99 77
5 KOAc rt 16 AcN 93 78
5 LiOAc rt 16 DCM 95 78
5 LiOAc rt 16 tol 95 74
5 LiOAc rt 16 ether 99 73
10 LiOAc 0 24 AcN 78 74
10 NaOAc 0 24 AcN 76 76
a
determined by 1
H-NMR, using 1,3,5-Tri-t-
butylbenzene as internal standard.
Table 10: Allylic substitution reaction using the dichloropalladium(II) complex of 5.
3.5 Summary
The synthesis of enantiomerically enriched PSiP (3) and PPP (4) pincer-like ligand was in-
vestigated. Difficulties were encountered due to oxygen transfer from sulfoxide to the elec-
trophiles used in the synthetic route. Nonetheless, an SPS type ligand 5 was synthesised as a
step towards the PPP pincer-like ligand 4. Complexation of the sulfoxophosphine ligand 5 to
palladium and plantinum was demonstrated and the resulting complex was carefully investi-
gated. Evidence of complexation to rhodium was found. The resulting complexes were tested
in asymmetric catalysis and showed moderate to good enantioselectivity.
49

70. 3 Synthetic approaches towards a chiral PSiP-Pincer ligand
50